Friend of mine is considering a MyECU for his Highland bike. We're trying to establish if it's possible / what is needed.

What are the sensor requirements for a P8? IIRC the tdc sensor sits on the flywheel, is that right? So is it waste spark or what? Or is that the speed sensor.

12-16 ohm injectors are no problem as far as I can see.

You don't support temp sensors having different specs, do you? I guess if they are, it's easiest to replace one or both of the sensors so they have the same spec.

Quote:

IIRC the tdc sensor sits on the flywheel, is that right? So is it waste spark or what?

Yes.
IIRC, yes.

No, I just found it it good old efiman.pdf. Or is that describing a P7 ECU? Flywheel has four tabs, one each 90 degrees and that sensor is just detecting speed. Another sensor sits at camshaft and senses phase. I'm not sure if that sensor reads one "tooth" or several.

Apparently the Highland just has a missing teeth wheel (60-2) at the crank shaft. After starting it doesn't know phase so it runs wasted spark until the ECU senses acceleration after each spark and determines phase. Not really following the KISS principle  :o

I guess we'll have to add some kind of phase sensor. What are the requirements for this in MyP8, missing teeth? Configurable?

Hm, if we go for MyP8 we'll need the external ignition coil drivers too. What are those? Easily obtainable?

You`re right, sorry, mixed up the sensors, I shouldn`t trust my memory that much... ::)
P7 & P8 same sensors.

raz wrote on 08/15/09 at 20:40:44:

Hm, if we go for MyP8 we'll need the external ignition coil drivers too. What are those? Easily obtainable?

Also the MyP8 is able to trigger the coils itself. Read the manual ;)
The temp-sensors are of course adjustable, look into the map ;)
Or am I missing something? I know someone who replaced the coil transistors with costly ultra low resistance FETs to address a similar problem

Why don't you stick to a wasted spark system? Do you want the MyEcu for spark only or also for a EFI conversion? I'd try even an EFI conversion with a wasted spark setup. Let it inject only half of what you need but this two times for one cycle.

I'm still planning to reactivate a SR500 and maybe convert it to EFI using a MyEcu. Then I'll probably end up with a similar system.

Hubert

On second thought: what happens when the plug fires into the open ducts while the first half of the charge is being injected?

A number of observations

first wasted spark should not really be the way to go 'cos it is just that 'wasted' and I thought the coils were designed to function with a specific optimum

second the idea is to replace the original ecu in total for something that is more accessible for modification, tuning using lamda and replacement parts.  My ECU seems to do just that

third I thought the P8, the My16m and the My 15m had similar inputs - is there a reason to pick one over the other?

Looks like this is going to be interesting

Well, my thoughts were: what's a wasted spark compared to a lot of development / time to get such an acceleration recognition coded?
The idea of sensing the crankshaft acceleration right after spark to recognise spark TDC is a nice one, looks very elegant. What I wonder is how they manage the injection thing while the ecu is wasting sparks right after start up.

Has anyone looked up their webpage?

Hubert

From what I can tell wasted spark is not really an option. This is not a carb-to-injection conversion but a crappy-efi-to-myecu one and I tell ya this Swedish Stainless guy is just like me in that he will just love to fiddle with injection phasing and stuff like that.

As far as I can tell the model will depend on sensor requirements. We have a crank shaft mounted, missing teeth wheel. Space constraints stop us from using a My15/16M type wheel but some kind of camshaft TDC sensor should be possible. My guess is the shortest route is to make some kind of TDC sensor on the cam shaft and go for a MyP8 - why not the mini P8 advertised somewhere here. Question remains: what are the requirements for this TDC sensor?

raz wrote on 08/15/09 at 20:27:01:

Apparently the Highland just has a missing teeth wheel (60-2) at the crank shaft. After starting it doesn't know phase so it runs wasted spark until the ECU senses acceleration after each spark and determines phase.

Sounds like a My16M or 15M is the way to go.
The Mk1 used a Centronics connector so it may be the easiest for custom use.
FYI: I'm doing a MyECU for the BMW twins. It uses a digital pickup like the Dyna ignitions.

Luhbo wrote on 08/16/09 at 00:43:49:

Also the MyP8 is able to trigger the coils itself. Read the manual ;) The temp-sensors are of course adjustable, look into the map ;)

Aha... so we just flick the DIP switch and use the P8 outputs as is? That is good news.

We have means to set the temp sensor values, but only once for both. What I meant was if the present sensors doesn't have the same spec we will have to replace one or both so they read the same. This is probably the least of our problems though.

_Cliff_ wrote on 08/16/09 at 08:08:47:

Sounds like a My16M or 15M is the way to go.

Please elaborate. You don't mean we can use the present sensor as is, do you?

Does the MyP8 use both "turning" sensors?

I don't see any reason why the current sensor can't be used.
Yes the MyP8 uses both sensors?

"Turning" sensors should have been "both rotation sensors, flywheel and camshaft".

Raz is talking about the phase sensor, I assume. The problem with the actualy mounted one might be that it is sensing the crankshaft, thus running twice as fast as the MyEcu normaly expects.

Hubert

Luhbo wrote on 08/16/09 at 17:56:39:

Raz is talking about the phase sensor, I assume. The problem with the actualy mounted one might be that it is sensing the crankshaft, thus running twice as fast as the MyEcu normaly expects.

Yes, and the ECU will have to run wasted spark as it has no means to detect phase, unless it calculates it like the OEM.

This subject beyond my knowledge but,

isn`t it possible to alter the values (configuration) in first row of a My15/16:
MyECU Cfg TDC=0x1600, CylOffs=270, PulsesPer=48, MissingPulses=3 
to fit the 60-2 wheel?

greenmonster wrote on 08/16/09 at 19:38:48:

isn`t it possible to alter the values (configuration) in first row of a My15/16: MyECU Cfg TDC=0x1600, CylOffs=270, PulsesPer=48, MissingPulses=3  to fit the 60-2 wheel?

I expect that to work but the ECU will still not know the phase. It will know TDC but it wont know if it's compression stroke or exhaust stroke. This is not a configuration issue, we need one more sensor.

http://www.weetwood.force9.co.uk/timing.jpg

What if you put a ring on this 60-2 crank wheel, with 4 small mags below it, sitting under wholes in this ring, and and add a second wheel somewhere on the camshaft/valvetrain according to the picture supplied by John?

I don't know the exact function of the phase sensor. As long as it just helps identifying the spark TDC out of the passing 8 (2x4, crankshaft runs double speed) teeth under sensor 1 (correct TDC if crank tooth AND cam tooth) you could also use a mechanic switch if this was maybe easier to be set up. 

Hubert

This is part of the discussion I had with Johan 'FettSnok'  our friend who is looking to replace the ecu

'I think I would be very tempted to pick up from the normal flywheel sensor on the Highland (I think the sensor is identical to the Guz) then weld in the missing teeth to give a full 50 or what have you in the signal. Then look at putting a single 'pip' on a cam drive wheel, sensed by another pick up.  The Pickups could then be wired in series to give a signal approaching that required. '

I'm sure you guys will sort something  [smiley=laugh.gif]

http://www.weetwood.force9.co.uk/cam.jpg

http://www.weetwood.force9.co.uk/cam1.jpg

http://www.weetwood.force9.co.uk/sensor.jpg

So, again, what does an ordinary P8 phase sensor send? Just one pulse, at TDC of cylinder 1, or what? We are stuck until we know what MyP8 expects from it.

raz wrote on 08/17/09 at 08:30:21:

So, again, what does an ordinary P8 phase sensor send? Just one pulse, at TDC of cylinder 1, or what? We are stuck until we know what MyP8 expects from it.

Come on Cliff - give us a steer on this.  Your 'How it works section'    http://www.cajinnovations.com/MyECU/howitworks.htm
shows something that seems to be a little different so I don't understand.   :-/

The pulse from the cam does not need to be accurate. It just needs to send a pulse before the TDC crank pulse but after the previous.

I think you should work with the orginal sensor and modify the map as suggested by greenmonster

john in leeds wrote on 08/17/09 at 09:01:58:

Come on Cliff - give us a steer on this.  Your 'How it works section'    http://www.cajinnovations.com/MyECU/howitworks.htm shows something that seems to be a little different so I don't understand.   :-/

That description was written for the original My16M. It is also true for the My15M.

_Cliff_ wrote on 08/17/09 at 09:07:21:

The pulse from the cam does not need to be accurate. It just needs to send a pulse before the TDC crank pulse but after the previous.

Actually it does need to be the TDC pulse. It can be any pulse.

_Cliff_ wrote on 08/17/09 at 09:07:21:

The pulse from the cam does not need to be accurate. It just needs to send a pulse before the TDC crank pulse but after the previous. I think you should work with the orginal sensor and modify the map as suggested by greenmonster

Ok then we know enough to make a decision. Creating that pulse sounds easy enough.

Do you mean you think wasted spark is the easiest route? What would be the drawbacks? Can a 15/16M be configured to divide the rpm by two to give a correct rpm reading? The injection will be pulsed twice with half the length so injector dead time will be more important to get right. Injection phasing can't be tuned.
Maybe it isn't such a bad idea, but that can't be done with a P8, right? So we don't have the option to start with this sensor and wasted spark, and later add a phase sensor unless we add a complete 15/16M type cam shaft wheel. Is that correct?

What Cliff writes looks easy. Stick to the actual sensor (a) and (b) add a second 1-cam sensor somewhere at the camshaft. this second sensor needs not to be in a very accurate position, it's enough if it comes a little before TDC of the main sensor. Its only purpose is to tell the ecu that the next TDC signal is the right one. Or maybe the wrong one. Take the wrong one and start counting up to 50 e.g. and then fire. Something like that.

John posted very detailed pictures. I think it's all in them.

Hubert

raz wrote on 08/17/09 at 15:49:19:

Do you mean you think wasted spark is the easiest route? What would be the drawbacks? Can a 15/16M be configured to divide the rpm by two to give a correct rpm reading? The injection will be pulsed twice with half the length so injector dead time will be more important to get right. Injection phasing can't be tuned.

It appears that all BMW twins work this way.It may be possible to determine phase it the engine configuration allows.

I don't know how the new beemers work, but the old ones not really were running wasted spark. It has to do with the 180° configuration. They had 1 contact breaker only, but due to physical effects only 1 spark was generated each time the points opened. The plug under compression usually ignites first, and once the ion channel is established all the energy is unloaded through this tunnel. This explanation comes close at least. This concept will not work for other engines (nevertheless you can see it on some badly done twin plug Guzzis from time to time).

And those old BMWs had carburetors. One problem less.

What's speaking against a P8? I still think this would be the option wich would suit this engine the best way.

Hubert

Please excuse me for being thick but could you tell me which MyECU uses both the cam and crankshaft sensors?  Are there separate inputs for these?

From what I understand Cliff to say

as long as there is a camshaft input (or pulse) then that can be used to fix the position in the engine cycle.

Other input sensor values eg engine temperature can be used but MyECU can modify them to give appropriate correction

My15 and 16M can drive both spark coils and injectors without additional drivers

Finally if My15M or My16M ecu take only one engine speed and postion sensor input from the camshaft can 2 engine speed sensors (1 from the crank and 1 from the cam) be connected in such a way as to produce a signal that MyECU can properly use?

Sorry for my ignorance guys, if you could answer my questions in order I am very willing to learn  :)

Luhbo wrote on 08/17/09 at 21:04:33:

What's speaking against a P8? I still think this would be the option wich would suit this engine the best way.

I think so too. Just add a simple phase sensor, one pulse per revolution, and even non-critical timing. Sounds perfect to me.

Now, if we use a P8 and add a phase sensor, do we have to "weld in the missing teeth" on the crank sensor as Johan described, or will it work fine as is? Or would he even have to rip all teeth off except four?

john in leeds wrote on 08/17/09 at 22:30:25:

Please excuse me for being thick but could you tell me which MyECU uses both the cam and crankshaft sensors?  Are there separate inputs for these?

Yes, the old P8 works as described in efiman.pdf which you quoted yourself.

raz wrote on 08/18/09 at 01:50:40:

...if we use a P8 and add a phase sensor, do we have to "weld in the missing teeth" on the crank sensor .....

I would not, because this wheel is well and exactly done. TDC is TDC on this wheel, so to say.
But, it all depends on how the P8 firmware works.

Another thing: what you said how the OEM ecu works is sounding so elegant that I expect this idea to be patented. You can copy it for your persoal use (who cares), but one can't copy and sell it then.

Hubert

Luhbo wrote on 08/18/09 at 08:05:45:

Another thing: what you said how the OEM ecu works is sounding so elegant that I expect this idea to be patented. You can copy it for your persoal use (who cares), but one can't copy and sell it then. Hubert

Maybe it's patented but it's not too uncommon. Aprilia has done the same. SAAB Trionic does something similar but more complicated, do you remember that document on ion-sensing during running? It describes the engine starting running wasted spark.

Apparently Sagem tried a system that just takes a guess on which is compression stroke, then tries to save engine from dying if it was the wrong guess. As you would expect it did not work too well :-)

The reasons I wouldn't be trying to fit a P8 to something that doesn't actually use it are

1) Its getting rare to get hold of.
2) Its getting expensive to get hold of.
3) Its f*ing huge.
4) It has 2 sensors

Other than that, go for it.

_Cliff_ wrote on 08/18/09 at 11:45:39:

The reasons I wouldn't be trying to fit a P8 to something that doesn't actually use it are 1) Its getting rare to get hold of. 2) Its getting expensive to get hold of. 3) Its f*ing huge. 4) It has 2 sensors Other than that, go for it.

What? When I say P8 I mean a MyP8 of course. I thought that was obvious. That invalidates point 1-3 and point 4 is given as we want to avoid wasted spark. That is also why we ask these questions here. Can we use the crank shaft wheel as is and add a phase sensor? Or do we have to modify the crank shaft wheel?

I know you mean MyP8. I cant make a MyP8 without a P8 and it is the same size.

Why don't you wait to see if wasted spark is an issue.
It may be possible to eliminate the wasted spark using just the crank trigger. If not, add the cam sensor.

Hi all!

It's me together with John in Leeds who are attempting to "remove" our stock ECU on the Highland.

Things to take into matter perhaps;

Why I want non wasted spark is that the charging time for the coils will be cut in half. The Highland uses separate COP:s, taken from some standard car. At 10000rpm each coil will ignite 10000times. Without wasted spark it'll be half. More charging time equals more energy to the spark, correct me if I'm wrong please. At compression ratios around 12:1 on some applications, I recon high energy in the spark is a key issue.

In order to determine if the engine is in compression or exhaust phase, a sensor for cam-position must be a must have? Can't see how to detect it otherwise, using only a crank trigger.  (If not doing as Saab, making an ion-read or as the stock Highland ECu using a flywheel acceleration measurement. Think either needs loads of coding/parts..)

Thanks for all your help!

And what about the injection timing? Don't know whether wasting a spark really hurts, but wasting a fuel charge for sure is no option. Dividing it into two also is not that easy. Raz wrote about dead time already, let me add the question when the injections should take place.

Hubert

BTW, P8 housings should be around in large numbers coming out of Fiats and other cages of that sort. What about the Mini-P8?

There's nothing other than the casing and connector needed from a WM P8, is it?

I was thinking a mini-P8, maybe in a completely custom casing, and preferably with Centronics type connector or some other easily available type. He will have to modify the harness heavily anyway.

I have no objections to using a My15/16M and then adding means for detecting phase, except I assumed it will mean custom modifications to both hardware and software and that is not quite as good as a "standard" MyP8 for a couple of reasons. MyECU is rare enough as it is  :)

I don't think wasted spark is an issue for spark energy. In a car application, they are firing 4 or 6 times more than on a bike with a single cylinder per coil. That's easy on the coil.

Depending on the engine configuration, it may be possible with all those teeth on the crank to detect when it actually fires and not.

The mini-MyP8 will not drive coils directly like the MyP8. You need the external ignition modules or you could always add a RecIgnition :)

[quote author=6F40454A4A2C0 link=1250279113/41#41 date=1250633576]

Depending on the engine configuration, it may be possible with all those teeth on the crank to detect when it actually fires and not.

quote]

So how would that work Cliff?

Sorry for insisting, but just for curiosity reasons: how does wasted spark and fuel injection work together?

We're talking of wasted spark not because someone likes to save some weight by getting rid of one coil or the distributor. Wasted spark came into discussion because in this special case the TDC sensor is mounted at the crankshaft and not as usual at the camshaft.

So every time you have a spark wasted you also have an injection wasted. This may be a good basis for silly jokes maybe, but probably not for a well running engine.

How do you guys think of this? I still could read no answer for this here.

Hubert

john in leeds wrote on 08/19/09 at 18:39:41:

So how would that work Cliff?

I expect the ecu would be able to detect an accelerating crank after a real firing.

Luhbo wrote on 08/19/09 at 18:47:54:

So every time you have a spark wasted you also have an injection wasted. Hubert

There's no such thing as a wasted injection. What you would have is a distributed injection.

Ok, maybe the wrong term. Distributed on two portions. This will work fine? Injector deadtime, valve overlap, injection advance, spark into valve overlap with fresh mixture hanging around in the ducts, resonance occurencies in the outlet concerning injection advance at certain rpms etc.

That's what comes to my mind when I hear "wasted spark".

Hubert

There is claims that spraying fuel on closed valve or open valve is a matter of rpm:s. At low to mid spryaing at an open one gives better mixture and then advancing the injection time to spray on cloesd valve before opening at higher revs. Injection advance is a real neat feature I think.

The casing is for me, not very important. An allround casing and whatever connectors that are available would do just fine I think? The harness will be built from scratch to fit perfectly anyway.

The key is still "how to detect ignition TDC" I think?

Can't say for sure but there might be some stock coils with internal drivers. John might be able to fill in here as he has these on his bike. I think there are three pins. Might be that two of them are power (+, -) and the last one is "trigger".
If this is a fact, then ignition drivers wouldn't be needed.  :-?

Edit: I think these types of coils are called "active coils"?
Think they need a logic 5v signal only from ecu?

Ok, what I wanted to say is: in my map injection advance is set to 520°. How do you transfer this to a EFI system which is running wasted spark?
Most if not all definitions like Injection Advance are based on the assumption that the EFI part is running with half the speed of the crankshaft.

Hubert

_Cliff_ wrote on 08/19/09 at 19:47:59:

john in leeds wrote on 08/19/09 at 18:39:41: So how would that work Cliff? I expect the ecu would be able to detect an accelerating crank after a real firing.

That is the existing highland system, it seems to work well on my bike now it has the twin inlet - starting is nearly instantaneous.  Does that mean a lot of work for Cliff though (how much will it cost us)?  It would be a good solution.

Luhbo wrote on 08/19/09 at 20:01:45:

(...) Injector deadtime, valve overlap, injection advance, spark into valve overlap with fresh mixture hanging around in the ducts (...)

Injector deadtime will be twice as important to get right, apart from that it is not a problem. But some people claim that spark at valve overlap is the cause of more than one carbed Guzzi catching fire and burning down to a pile of crap. Right or wrong, I have no idea. Carefully choosing the injection advance should decrease the risk.


Luhbo wrote on 08/19/09 at 21:38:59:

Ok, what I wanted to say is: in my map injection advance is set to 520°. How do you transfer this to a EFI system which is running wasted spark?

While I dislike the wasted spark idea in theory (or for purity) I must admit I don't think it will be much of a problem, neither spark or fueling. At least not until you want to squeeze the absolute last hidden powers from the engine, completely impossible without a good brake dyno and lots of invested time.

Your 520° [trailing edge] is a very wierd figure as it is, if we do some timing maths. At idle it will inject the whole shot on closed valve, but as pulsewidths get longer and rpm increases, it ends up injecting some of it all the way back in the previous cycle's intake stroke and the rest at closed valve. How wierd is that? Actually this will be the case as soon as at 3500 rpm if you give it a handfull of throttle.

Running the ECU in 'two-stroke' mode as some call it, let's say we use the corresponding value 260°. This means we'll get the first shot of the two injections at closed valve and the other one ending at maximum valve lift. I wouldn't be too surprised if that even turns out to be of benefit.

GreaseSnake wrote on 08/19/09 at 20:41:35:

Can't say for sure but there might be some stock coils with internal drivers. John might be able to fill in here as he has these on his bike. I think there are three pins. Might be that two of them are power (+, -) and the last one is "trigger". If this is a fact, then ignition drivers wouldn't be needed.  :-? Edit: I think these types of coils are called "active coils"? Think they need a logic 5v signal only from ecu?

Hi Johan,

the later ('04) coils are indeed 'active' see picture.  I include the wiring diagrams for both '01 and '04 and onwards models

http://www.weetwood.force9.co.uk/coil.jpg

Wiring diagrams for '01 models - shows coils are not active

http://www.weetwood.force9.co.uk/01eng.jpg

http://www.weetwood.force9.co.uk/01ign.jpg


Wiring diagram for '04 and on showing 'active' coils

http://www.weetwood.force9.co.uk/04eng.jpg

http://www.weetwood.force9.co.uk/04ign.jpg

BTW I have a spare (no longer working) late pattern ecu

raz wrote on 08/20/09 at 02:54:12:

... Your 520° [trailing edge] is a very wierd figure as it is, if we do some timing maths....

In fact it's not weird, it's just a dilemma. The charged fuel needs some time to evaporate, finding this time while laying around in front of the closed valve. A dilemma I call it because it's obvious that with higher rpms and wider throttle this time gets shorter and shorter. In fact it is the case that at WOT you'd need the complete two crank revolutions to get the required amount of fuel completely delivered. Lambda 0.8, original airbox and open Mistral cans can push you so far ;) In this case you'll have to face other problems arising.

That's one of the unsolved problems btw. which high reving direct injections engines have to deal with.

And thanks for puting in the expression "Two Stroke Mode". It surely helps to keep the discussion focused. I don't know whether I would have posted again if the discussion whether a spark should be wasted or not would have went on  ;)

Hubert

john in leeds wrote on 08/20/09 at 03:49:59:

GreaseSnake wrote on 08/19/09 at 20:41:35: Can't say for sure but there might be some stock coils with internal drivers. John might be able to fill in here as he has these on his bike. I think there are three pins. Might be that two of them are power (+, -) and the last one is "trigger". If this is a fact, then ignition drivers wouldn't be needed.  :-? Edit: I think these types of coils are called "active coils"? Think they need a logic 5v signal only from ecu? the later ('04) coils are indeed 'active' see picture.  I include the wiring diagrams for both '01 and '04 and onwards models

So the 2004 version have four pins:
1. signal from ECU (nothing tells us polarity or voltage)
2. not connected
3. ground
4. this pin is interconnected between the two coils but otherwise it goes nowhere except to that CAN bus connector

They don't seem to have an own +12 V feed. If that is correct, they can't be active ones I guess.

Quote:

from Raz So the 2004 version have four pins: 1. signal from ECU (nothing tells us polarity or voltage) 2. not connected 3. ground 4. this pin is interconnected between the two coils but otherwise it goes nowhere except to that CAN bus connector They don't seem to have an own +12 V feed. If that is correct, they can't be active ones I guess.

Dammit, screwed up again  ;D

Obviously a little knowledge is a dangerous thing

_Cliff_ wrote on 08/19/09 at 19:47:59:

john in leeds wrote on 08/19/09 at 18:39:41: So how would that work Cliff? I expect the ecu would be able to detect an accelerating crank after a real firing.

Could the event that sets the phase counter just be the movement to the next level of the speed after cranking?  Cranking giving a lost spark and 2 injector openings.  Moving to the next engine speed level after the cylinder fires (the engine accelerates) the phase point is set and normal service resumes.

As I said before a little knowledge can be a dangerous thing so 'just sayin'

Probably not. To vague I think. Just think of an humble start for instance.
I'd set the ECU into "Start Mode" after power up or better after it detects zero RPM (no pulses). In this mode the processor does not only count the phase pulses (the teeth) but also measures the time between them. In case the processor sends a Spark signal AND detects more or less simultaneously a sudden and remarkably shorter time between the pulses it might have been the spark TDC. On the other hand if time between pulses becomes longer then the engine is probably nearing a compression TDC.
As soon as the RPM has jumped to the second row the program falls back into normal running mode and behaves as usual, as today.

Hubert