Bmw r75/5 microsquirt project for a newbie

[fefenin]

Hello everyone,

I m reading the data sheet sinces few days while trying every single function on bench.

I plugged every sensor except IAT and the coils.
Everything looks like ok.
But I don't know how to start with the pickup sensor.
I would like to get rid of the old trigger points that are mounted on a mechanical auto advance wheel.

I baught with microsquirt a 4 wheel but it doesn't fit in the engine. So I wonder if there is simple other solutions cause the VR trigger is not a good option . I would préfère optical or hall sensor.

So my questions would be:
Is it possible to use the missing tooth wheel downscale with an optical sensor?

Is it possible to glue two magnet 180 degrees apart on the original wheel to use the hall sensor?

Last question is when I tried to crank up my fake wheel I discovered that only one injector is driven (inj1)
When I set my engine to even fire.

I thought a flat twin boxer like would be even fire?

Am I missing something?

Thanks in advance

I'm so excited I hope I can make this project comes to life it is a young kid dream.

When ignition works then i have everything from an old 748 Ducati to make it go injection.

[LAV1000]

This one is also beemer just no R75
Take a beer or coffee and start to read:
http://www.msextra.com/forums/viewtopic ... 01&t=57652

[R100RT]

Perhaps I can help you on your project. I have the same basic engine arrangement with my project (1983 R100RT) & running successfully now for many years.
Post some photos of the parts you have already purchased.
You mention a "4 Wheel" is that four teeth and where are you mounting it on the engine?
VR sensors actually work quite nicely with Microsquirt & I have one each for crank and cam sensing.
Another thing that will be a big benefit for you is a "Jim Stim" device to test everything off the bike etc.
Let see some photos.
Regards, Lorne

[SQLGUY]

I have the same question, I think: why do you say a VR sensor is not a good option?

VR was stock on my '82 Yamaha SECA 750. I fabricated a 36-1 wheel to mount in place of the original crank wheel and the stock sensor was fine (after some mods to the original uS conditioner circuit), and, it's my impression, at least, that I get quicker synch and better timing accuracy with the 36-1 wheel than I did with previous 1 tooth, and 8-1 tooth wheels that I tried.

Anyway, can you post a photo or two of the crank-end, with some details about why mounting a VR sensor there wouldn't work?

Some wheel variants I tried:

IMG_0238-small.jpg

And how it looks on the bike:

SECA36-1Wheel2.JPG

[R100RT]

737.jpg

Similar 36/1 wheel on the end of the crank nose (alternator has been removed and is a modified drive for larger item.) That's a common Ford VR sensor - can be found in any parts store or junk yard for peanuts & works reliably.

DSC02187.jpg

This is a modified bean can (your's perhaps is a points off the cam set up), shortened to clear the aftercooler and contains a single toothed wheel - for sync on sequential ignition. Again, a common VR sensor that is boringly reliable.
You will probably remain with wasted spark and won't need to condition the signal.
Speaking of charging systems, your bike will have modest electrical reserves that can be overwhelmed by addition of fuel pump, and other demands. Some LED light upgrades might be in order to reduce consumption levels.

[fefenin]

Oh thanks guys . I m an electronics so logic levels sounds better but I did put the 36/1 wheel on a computer fan to try it on bench and it is working very well.
Now the thing is I would like to keep my alternator so I would have to downscale quite a bit the wheel so it fit in the engine crankcase. I ll show you some pictures tomorrow. It is close to midnight here.

I also made the spark external circuit and it works great but only Inj1 and ignition1 are triggered. I don't get why?

Plus I will need your help to set the right timing on the wheel according to the TDC.

Everyone's help is much appreciate.
I m not sure I'll sleep questions cannot stop in my head now!

Thanks a lot for now

[fefenin]

Oh and by the way what is the benefits to have a cam sensor as well as a crank sensor?

[rickb794]

Timed injection is possible with addition of a cam sensor

[R100RT]

As mentioned, a cam sensor will allow sequential injection & ignition. I referred to "wasted spark", and without knowing any details of your project offered that as the most likely ignition control strategy you will be using - based on it also being on our bikes already (twin 6v coils, wired in series and both fire together on each signal from your mechanical points, again a guess on current bike equipment). I can offer why I went with sequential ignition, after having initiated my forced induction project with wasted spark, but it is a controversial subject - so lets stick with your questions (please utilize the private message feature here if you wish to probe details on that or any other matters concerning our airhead engines and upgrading them).

The link provided by LAV1000 is for another project by a fellow named Jed - with a 1994 R100GS. I talked with Jed during his project and lent a Jim Stim for some troubleshooting - if you read that whole post you will note that at the very heart of the matter - some basic electrical rules were not followed and that led to problems with electrical interferrence. Your best chance of success then will be to follow the manuals that are listed in this area for the product you have purchased. Then, anything that isn't quite clear will be answered rapidly by those on the forums that are very helpful in stating the obvious and helping return things always to the simple truths and success .
Although the basic well maintained Bosch charging system on our bikes may have as high as 280watts at the higher rpm's for charging, that can be adequate providing power isn't wasted and riding style is aceptable (for example, a lot of low speed, city stop and go with accessories on will most likely bring your battery state of charge down from acceptable. In that case a battery minder when parked can offset problems).
For your tone wheel, there are smaller sizes that can work (Jed used one from Silent Hektik I believe, with less teeth). 36/1 has very good resolution however. A tone wheel can be piggyback mounted onto the Bosch alternator where the retaining allen bolt secures the rotor, the front cover can be spaced forward about 8mm with a hand made spacer (I prefer Aluminum) to provide better clearance for a wheel with more teeth.
Regarding the incorrect behaviour of your testing, you will gain much better response by posting an MSQ and short recorded logs for others to see the details

Welcome aboard, always exciting when another airhead surfaces in the MS world here.
Lorne

[fefenin]

hello,

thank's for this answer... i thaught i could go sequentiel injection without cam sensor (not sequential ignition)
this is the part not clear for me , (my mother language is not english) :

why can't we get every information necessary for a sequential injection on the cam sensor with 32 tooth?
does that means that the decati i took the parts from was not sequential caus i goit only the tooth wheel plus one vr sensor?

here i list what i've got:

-tps from ducati 748 i would say / tps works and is calibrated (the whole assembly 2 injector plus 1 tps + thr body looks like a 2001 monster s4)
-map from ducati was only 100kpa so i swap for a bmw525tds map (seams to work but not sure if it is well calibrated)
-clt from ducati 748 up an running
-vr from ducati 748 + 32/1 tooth from autotune (i'v got the ducati wheel also that is way heavier and bigger)
-narrow band lambda + gauge but not plugged yet, works good as a standalone sensor.
-coils from ducati 748 plus external bip373 driver kit . up and running
-injectors from the ducati thr body up and running
-i forget two knock sensor that i read i need something for the signal to be right for microsquirt


msq is on but for the log i'm not sure what you need...

so my question now is howcome i get only one spark and one injector fired?
is it because i selected sequential and even engine and i have only one pickup (no cam sensor)?

and yes private messaging would be good so i know the option i've got from sequential to wasted spark plus keep or not the charging system.

[hybrid]

It is not possible to know which stroke the crank is on when you only have a crank wheel.
Remember the crank will go past TDC both times piston #1 is at the top, but it is only on the compression stroke one of those times.
So your crank wheel will also go around twice each time, but only once is it in the right place for properly timed injection and spark.
The extra cam sensor tells the ECU when the engine is on the compression stroke.

[fefenin]

Make sens now. But this engine is working with wasted spark anyways at first. And it might be not so critical that the squirt occurs of the timing , gaz will seat behind the valve right?

[fefenin]

Can I go wasted spark and squirt with the sensor on the cam then? Like that I don't need to mod or destroy the power supply generator .

If it works that would be a good start for me to familiarize with microsquirt settings

[R100RT]

OK, you have the small cam driven assembly with the flyweights for advance on the original points driven ignition. This provides two signals to the coils per crankshaft revolution. If you wish to utilize a signal from that you have a couple of things to be aware of:
a) The resolution of the cam based signal is rough (chain bounce, flyweight plate jiggle, mechanical factors on the points and cam lobes).
b) You have advance in the ignition map, so that spring/ weight assembly is counterproductive.
c) This is not something I have tested but I believe others have found it problematic and will lead to sync loss and resets - in other words bad performance. That may have been good enough to run the orignal coils but confuses the MS and resets occur.
d) You would be best testing the signal while it is connected to your Microsquirt and view the composite tooth/ trigger logs on Tuner Studio. You should have the advanced Tuner Studio to get this project running correctly and it is very reasonable. I would try to mimic the operation of the points (if that is the wish to use) with a micro switch on a variable speed drill for instance, or a timer relay that can be dialed for contact opening/ closing through a variety of speeds (I have an Omron relay for instance that can be set to cycle so fast that it acts in miliseconds and I use it for tests like this) In this way you will witness if your Micro Squirt will accept that low res signal from the low cranking rpms. Any testing on the bench before the bike is started is going to be your most productive time spent.
e) You have a 36/1 wheel from DIYA I assume, that is the best option when run off the crankshaft, with a VR sensor for instance. A small adaptor could be machined to fit to the end of the Bosch alternator rotor, etc. etc. A spacer for the front cover would be required. Or, utilize a smaller wheel with less teeth (less resolution however) to reduce amount of space required.

You are fine with wasted spark. The difference in the way the bike will perform if sequential is used will be hard to notice. Add a turbocharger and I would then strongly recomend otherwise, but others may disagree.
There is another method to gain sync for sequential ignition/ fuel when a cam sensor is not used, you will read of it in the manual - utilizing the map sensor to detect phase from compression/ slow down during initial cranking. A dual wheel with missing tooth is choosen under "Trigger Wheel Arrangement", and under "Cam Input" you choose map sensor. Then, the map sensor phase threshold is set under General Settings Screen (this is describe in section 8.2 of the Tuner Studio manual.) I have not tested this arrangement, again you should be fine with wasted spark and no phase detection.
You will have to go back in and change your sequential injection to "untimed" I believe at this point. Your lags under general settings are all quite tight and you might find leaving them towards 100 till you've run and tested would be better to make corrections there.
Tuning with a narrow band will not be so easy, why not a wide band (the best info).
You may have some challenges gaining tach signal for your bike tachometer but there are a few tricks that can be applied there once you've made it that far.
First and main thing is the ignition signal and a robust set up. I wouldn't recomend relying on the original points.

[fefenin]

ok fine , you conviced me the flywheel with missing teeth is the best option on the cranckshaft.

so i will go untimed and with the vr sensor with flywheel but don't you thinq i can use this wheel plus sensor on the cam? so it fit with to much mods cause i have no machines left to do trhe work...

i would go for wide band but after a while and first i would like the ignition to work not the injection...

my test setup showed it work with the 36/1 wheel from DIYA but i have to put odd fire and put 180 degrees in odd fire first angle for it to work properly and trigger injectors and sparks.

can you comfirm it woul work that way?

for the software and the lambad sensor i will get it as soon as i got money but this little ecu took all that i've got for now.

how do you mesure on the advance timing ? can you set up when to trigger the spark accuratly? the original mechanical system is set to 6° for cranking i guess

[R100RT]

If you intend to run the 36/1 wheel on the cam nose, I am not familiar with the settings that might allow that to work - perhaps others will chime in to comment.
However, from a mechanical standpoint the end of your cam is a rather fragile & miniature shaft that I don't think will support the rigors of forces from a fairly large wheel mounted to it, imbalance and runnout that those are famous for (in fact, that creates a slight timing imbalance between jugs on the stock operation where two cam lobes operate 180 apart and creates a timing light "double image" when viewed)
The later bean cannister with points or hall sensor had bearings in it and could support a bit more rotating weight (what I used for my single toothed sync portion). I believe Jed was testing that also on his project (the link via LAV1000) but he ended up choosing crankshaft mounted wheel and speed.
It may come down to just trying what you wish to experiement with, I gather you don't have access to machine tools although it would not take much for the crank adaptor - I'm kind of a broken record on that score.
For your ignition table, you can put your best estimates in place, as well there are numerous timing maps on the net for our engines that can be looked at - but most of those are based on compromises via equipment that isn't as flexible as the MS, and you are better fine tuning this on the dyno (to avoid damage) or very carefully via road tuning and extensive logging.
Keep us informed, more photos too.

PS - a link over to the ADV forums if you wish to view my project photos and such - http://advrider.com/forums/showthread.p ... R100RTurbo

[SQLGUY]

I'm not sure on the odd-fire offset. Your firing should be at 0 and 360 (minus advance) for one cylinder (the one that's at TDC wherever you have designated TDC for the 36-1 wheel and Microsquirt), and 180 and 540 for the other cylinder.

A timing light is what I used to verify timing. Etch a mark on the 36-1 wheel for TDC and check that relative to the VR pickup or some other fixed reference on the crankcase.

Edit: if you don't have a timing light with adjustable advance, you could glue one of these to the 36-1 wheel: https://www.google.com/url?sa=i&rct=j&q ... 5248235571

Also, that 36-1 wheel you showed looks much heavier than what's needed. The one I made with the square teeth was made from two pieces of electrical box steel plate welded together. It's about 3mm thick. 3 or 4 millimeters thick should be fine.

[fefenin]

@ R100rt . Ok I decided by reading the real manual , and because it is very well explained:
To use a home made 12-1 or 24-1 homemade wheel on cranckshaft. So it is lighter and it fits better in the case plus the manual days it is better for motorbikes and resolution should be ok.And a one teeth optical sensored on cam shaft
At first I was reading the quick starting guide (yes I thaught it was the full thing) and I didn't get all that I got in the proper manual.

I will probably never acces to a dyno it is not very common to mod anything in france.
So is there a way to safely try it out on the road?

As for machine I use to have a cnc, a lathe . Still left 3d printer that I can use for trial parts before a friend can make me the proper parts I guess.

Thank you for the link I'll have a look later when not on my phone.

@SQLGUY thank you for the input . By reading the data sheet it is clear how to set up your TDC and then count the angle and put that for tooth#1
At least it seams easy for crankshaft . Will see for camshaft

At everyone:
Is there any possibility that microsquirt is alowing only one bank of injector and ignitor when on 2 cylinder evenfire four strokes ?


Thank you guys for helping me . I don't think I can make it alone

[fefenin]

i found this just for info:

6.9.7 Missing tooth cam wheel
This arrangement is not commonly used by OEMs but does support full sequential with a single wheel and
sensor. Cam triggering is less accurate than crank triggering due to timing belt or chain stretch.
The software benefits from a reasonable number of teeth (hence 36 or 60) for best ignition timing accuracy. Low
tooth count wheels such as 8-1 are not advised.
The previous section on missing tooth crank wheel generally applies when the wheel is mounted to the cam, but
remember that one rotation of the cam is 720 crank degrees. The settings are in crank degrees. So a tooth#1
that is 8 gaps earlier than the sensor on a 36-1 wheel would give a 160deg tooth#1 angle (8 * 10 * 2 [for cam] )

as you said it is less accurate due to chain stretch and things but you can go sequential with one wheel on cam only

plus i did some printing so i can see if it works well , tomorrow i'll do it in metal anyways but i'll try full optical just for fun:

20160326_210410.jpg

how is your one tooth cam R100RT?
does it start when the missing tooth is coming on crankshaft?

[SQLGUY]

On my Seca I did dyno tuning originally, but, more recently, did street tuning. Street tuning is not too bad... some hills help, as do areas where you can drive fast, and slow, and accelerate quite a bit, without much traffic around you.

To do street tuning, though, you do need a wide band EGO, and a laptop.

Basically:

1. You set your VE map based on a best guess... there's are some tools that will create a starting map for you...
2. Then you set up your desired AFR map based on areas like idle, cruise, coast, and power/WOT
3. Then you can use Tuner Studio's auto-tune to adjust no-load areas of the map... which will give you an idea of how close (or not close) some areas of the guess map are. You can also adjust the whole map a bit based on that - like, if idle and no-load were 20% too rich, you might want to scale the whole table down by 15% or so (too rich is safer than too lean when starting tuning)
4. Then you can strap the laptop to the bike, start auto-tune, and ride around a bit under varying circumstances to auto-tune the cruise and power areas of the map.

Almost none of this works with a narrow band, because there are few areas of the map where you want perfect 14.7:1 AFR. Older engines usually want a bit richer than stoich for idle, for instance. You could compromise and tune your idle, low load, and cruise for 14.7:1, then try to guess richer for other areas of the map, and maybe enrich the idle a bit from there... but I wouldn't recommend it... a narrow band EGO isn't free either, and an LC-1 or LC-2 is not very expensive, and tuning even the low load areas will be faster and better with the wide band.

I did have one problem with the street tuning, where, I think maybe I had a reset (Microsquirt - mine, at least - seems to be pretty prone to resets when hooked up to a computer under load and higher RPMs; I don't have the same problem when running without the laptop), or Tuner Studio just lost its mind. I had done a couple of runs, hitting most of the map, and the tune was getting pretty good, when, all of a sudden I started hitting what felt like a rev limiter at 7K, then 6K, then 5K... When I got home, I looked at the map and found that Tuner Studio had significantly leaned out a large swath of my mid-band. I corrected the map, retuned, and TS did something similar, but not quite as extensive again. After one more clean up of the map and another rerun, I got good tuning in those areas, and, overall, the bike is running really well - great driveability under all conditions, great throttle response, and great power... right up to red line.