PDA

View Full Version : Magnetic sensors



MoTeC
12-11-2003, 07:49 PM
Hi Guys,
As you may have heard already the team from Germany had ECU/electrical problems all weekend. Unfortunately these could not be fixed with the fitment of an M4. http://fsae.com/groupee_common/emoticons/icon_frown.gif I thought I would give you guys some background so the problem does not surface again, it ruined their weekend it could ruin yours.

The car was fitted with a reworked factory ECU, Bosch I believe (or something similar). The ECU required a 36-2 trigger disc which is fairly normal for European stuff and a magnetic sensor. The ECU seemed to be firing fuel and ignition but not reliably enough to run the engine, this is where the N.S.W. guys stepped in with the offer of a spare M4. I was asked to wire the thing in and get it going, no problem the boss seems to pay me to do this sort of stuff.

The ECU was wired in, all the inputs and outputs checked, no worries. Now try to start the thing.....Hmmmm it comes up with a few errors....get the scope. Oh dear, one of the wrost crank trigger traces I have ever seen. Not because there was heaps of background noise but it was a trace you would ecpect from an engine that was not turning over at a constant rpm. Take the spark plugs out and check again....no change. Basically it looked like the teeth were not concentric with the crank and this was causing pluses that ranged from 2volts peak to peak to about 8.5volts peak to peak, that's pretty bad. Our ECU was picking this up as an extremely rapid change in RPM and causing errors that would not allow the ECU to start the engine either.

The old ECU went back on and luckily for me it suffered the same error. The interesting thing is that because it is based on a factory ECU there is a ton of contingencies for just getting the thing started to limp to the work shop, aftermarket ECU generally do not do this, they rely on a good signal. The car did a couple of laps but unfortunately that was it.

A quick bit of testing on the lathe back at the office gave me this:
With a trigger disc with a 145mm diameter and about 0.3mm out of concentricity (is that the right word) I scoped a pattern that had an average pulse of 3.6 volts at 300rpm the range of pulses was + or - 900mvolts around this, that is nearly one volt! Also on another test going from 0.5mm air gap to 1mm air gap halves the pulse (logical when you think about it).

Moral of the story: If you are going to retrofit different components to your engines make sure you check everything!!!! Another issue with the German car was that ther was not enough clearance between the chassis and the engine to get the trigger disc or the sensor out for modification. The other BIG downfall for me is that I have to explain to the boss why I could not get the car to run....not an easy thing to do. http://fsae.com/groupee_common/emoticons/icon_frown.gif

Hopefully the boys can get this sorted for the next event.

Mark McCoy/Donna Arbuckle

Charlie
12-11-2003, 09:27 PM
Hmm, sounds similar to reversed polarity on the sensor wires does it not? Thanks for the lessons learned,

-Charlie Ping
Auburn University FSAE (http://eng.auburn.edu/organizations/SAE/AUFSAE)

MoTeC
12-11-2003, 09:54 PM
Hi Charlie,
Not for us, we can tell the ECU which way the signal goes so you simply wire it which ever way is quickest. http://fsae.com/groupee_common/emoticons/icon_smile.gif

The rapid change in voltage at constant revs was the problem in this case, have not come across it before, it will go down in the book of 'wierd occurences'.

Mark McCoy/Donna Arbuckle

Eric
12-12-2003, 03:08 PM
What kind of magnetic sensor was it? I've seen sensors that are a switch based on hall-effect. If they see a target, they close. Otherwise, they are open. I have had trouble with the air gap opening too wide, and the sensor missing the tooth. Otherwise, I think this type of sensor would avoid the voltage variation.

MoTeC
12-12-2003, 05:07 PM
Hi Eric,
If the trigger dics is non-concentric you may miss some of the teeth altogether which would be just as bad. You have to take your mounting position into account as well. Most of the bike engines you guys are using have the trigger disc and sensor more or less monted in the crank case. Hall sensors and other electronic sensors do not like too much heat so a magnetic sensor is probably best in these situations. Also Hall and electronic sensors usually need bigger teeth than magnetic sensors (general statement) so trigger disc size can be a problem.

Mark McCoy/Donna Arbuckle

Tim Heinemann
12-13-2003, 04:14 AM
Speaking of Hall effect vs. induction sensors we had a problem on our first car with the cam position sensor we wanted to use in order to get sequential injection and ignition (hope this is the correct term). The problem was to get the engine started: A magnetic sensor did not reach a sufficient output for our ECU (a TriJekt, German aftermarekt product) because the camshaft did only 120-150rpm (obviously half the crankshaft speed), regardless of the air gap. As there was no Hall sensor at hand at the time we rewired for batch injection and wasted spark which gave a fairly good starting behavior (the engine did not run well because we never got the turbo running properly due to lack of time and following the GB event in 2002 it has not been touched as we're focussing on getting the Mahle engine running).
I wonder what most teams are running, batch or sequential injection and what you use as camshaft sensors? As soon as time permits we want to get the Suzuki-powered car running just for the fun of it and sequential injection would be much more interesting to play with IMHO.


Tim

MoTeC
12-14-2003, 09:22 PM
Hi Tim,
What you have to remember about magnetic sensors is that the signal is related to how fast the tooth passes and also to the number of windings there are in the sensor. The sensor I was testing yesterday had over 1volt peak to peak at 190 RPM so that is not too bad. and that is with a fairly small trigger disc.

One thing to try is have a longer tooth off the cam shaft that will make the surface (for want of a better word) speed of the tooth greater and therefore the signal out of the magnetic sensor bigger.

Sequential is the only way to go. http://fsae.com/groupee_common/emoticons/icon_smile.gif

Mark McCoy/Donna Arbuckle

A Reinke
12-16-2003, 08:48 AM
Mark/Donna - did this German team make the trigger wheel or is this a stock wheel? i'm assuming it was machined...horribly? since it is so out of round.

i do have to say Motec's allowing us to use the stock trigger wheel and sensor has made my job in getting the engine running much easier. the only tricky part was setting the degree offset from TDC to the gap in the wheel.

the wheel we modified/made to use on the old Electromotive ECU never worked...the M48's 'diversity' saved us.

~Adam

MoTeC
12-16-2003, 04:20 PM
Hi,
I could not see the trigger disc on the German car as we could not get to it. They made their own trigger disc. The 36-2 was the way the factory style ECU needed which luckily is no problem for us. The trigger would have been a fair bit smaller than normal to fit in the space, this could have been part of the problem as surface speed of the teeth is obviously going to be smaller.

We put a lot of R and D time into writing software to run factory setups, some of the new ones are really strange. http://fsae.com/groupee_common/emoticons/icon_smile.gif

The Germans idea is fine but the way it was mounted/machined wil need a bit of work.

Mark McCoy/Donna Arbuckle