Gen Board/Manual/Input Trigger Types

GenBoard/Manual/InputTriggerTypes (2006-07-29 04:29:43)

This page is in terrible need of organization and clarification. If you can help straighten it up, please do so.

Sensors and mounting:

The HALL sensor needs a magnet too. We should add these magnets to the WebShop (since they are available; should it be 1 Euro?):

dia=10mm, 5 mm tall
dia=15mm, 10mm tall
dia=18mm, 10mm tall

TLE4905:

if the magnet is on the rotating part (practical on the CAM since it's just 1 magnet, or on the crank for even-type trigger - naturally prohibitive for multitooth), it is very easy to mount reliably: max sense distance is 5..16 mm depending on the magnet size (for magnets above)
a fork-type works too: wheel between magnet and sensor
but magnet behind the sensor is not practical here. That setup requires a special, differential HALL sensor (which has 2 HALL sensors inside) or analog HALL sensor and using the onboard LM1815 VR chip. It is possible to make it work, but it is very sensitive, partly because of the considerable histeresys of the TLE4905 (which is otherwise good for preventing false triggering).

These already have the magnet:

On GenBoard/VerTwo you need to make a small connection for the input trigger on the board:

You need to connect the output signal of the opto (inside the board!) to the IC3 signal (pin 9 of the avr, available on 1x1 header J8). INT6 input was used originally, but we changed to use the IC3 pin for the primary trigger, because it has hardware-capture capability (more precision), and reserved INT6 for the secondary_trigger (CAM signal for dual-trigger applications, or measure alien ignition advance).

Look here: http://www.vems.hu/wiki/index.php?page=GenBoard%2FVerTwo%2FPCBLayouts

The wire needs to go from the R18/C13 junction to J8 (the pin closer to the AVR). The physical place on the board is just to the right of the AVR. I used the lead of R18 (close to C13) to solder the wire on, then soldered it to the J8 header. There is no need to cut the trace to AVR pin 8 unless you need to measure alien ignition. If you already have a pin header in the J8 location, it is NOT recommended to desolder it. Many times damage occurs from trying to desolder. If there is no pin header in that location, you can just solder the wire to the pad.

Both the primary_trigger and the secondary_trigger on GenBoard/VerThree can be configured (independently) to accept any commonly used type of trigger signal.

See InputTrigger/AnalogSide for the amazing wide range of the chosen LM1815 circuit.

Inputtrigger can use any technology as long as you can acquire a ttl level signal in the end:

mechanical (usually with pullup resistor)
optical
HALL type (very popular)
VR (variable reluctance)
the output of a full already installed system, like EDIS

It's important to have clean signal from the inputtrigger sensor,especially when using ignition.

See [ http://www.picotech.com/auto/waveforms/crankshaft_sensor.html various crankshaft sensor waveforms]

HALL

See MembersPage/MichaelKristensen/CamHallSensor for a custom HALLsensor installation.

To get a clean signal:

have the HALL sensor supply only from GenBoard
sharing/tapping other ground, especially chassis is very bad practice
Pullup from Hall output to Vcc R=10k (Genboard V3.0 has no place for this, connect it where it fits. V3.1 has R30 and R48 for this)
C between Vcc and Gnd, 1-10uF
take care of the magnet location. With the histeresys of the digital output sensor, bad mechanical mounting results in lost or poor signal. Check [Micronas HALL pages] for a good introduction to HALL sensors

This is good:

Look on MembersPage/ChrisGerhardt/ChrisNotes for his HALL wiring and noise experience. He went through a great deal of ground noise issues and with everyones help seems to have a very clean trigger now.

See [for distributor hall sensor waveform]

HALL sensors:

The commonly used HALL sensor is a unipolar digital output (with histeresys) switching at appr. B = 4..10 mT [TLE4905L], as in our WebShop [shop].

Developers, please check these out, which would be useful:

HALL SENSOR TLE4905L (INF) DIGITAL TEMP COMP UNIPO this is the most commonly used type: low B threshold, but relatively high histeresys.

HALL SENSOR SS443A (HWL) DIGITAL TEMP COMP UNIPOLA I have a few samples of these, just tell me if you wanna try (in WebShop order TLE4905L and write SS443A in the order note). Looks like higher operating B, but smaller histeresys (11.5 / 14.5 mT) than TLE4905. Guaranteed upto 150C.

HALL SENSOR SS461A (HWL) DIGITAL TEMP COMP BIPOLAR

HALL SENSOR SS49 (HWL) LINEAR NON TEMP COMP 4-10VD

HALL SENSOR SS495B (HWL) LINEAR TEMP COMP 4.5-10.5

HALL SENSOR SS94A1F (HWL) LINEAR TEMP COMP 6.6-12.

One more choise from Allegro: A3240 unipolar HALL, 4.2-24V [see more]

Optek OHB900, hall fork for vane type trigger wheels, http://www.elfa.se/pdf/73/730/07308406.pdf This was reported to be inconsistent at high temperatures. -JÃ¯Â¿Â½gen

http://content.honeywell.com/sensing/hss/SpeedAndPosition/o1-2.asp

Magnet and mounting

Forget Neodimium magnets in the engine bay, they become unreliable sligthly above 80C (!!!). Also, Neodimium magnets corrode as soon as the Ni cover is scratched. They are more expensive than simple ferrit magnets, and the extra strength is not needed. What is needed is good way of mounting. The easiest way to get a good signal is to have the magnet and the sensor on opposite sides of the (ferromagnetic, windowed material) wheel. However, this way of mounting is often non-trivial if space is scarce.

VR - Variable Reluctance

The output amplitude of a variable reluctance sensor depends on RPM. It will make appr. 100 times bigger signal at 6000 RPM than it makes at 60 RPM cranking. To suppress noise at high input levels while still allow to trigger on lower signals an adaptive hysteresis circuit is utilized using a peak detector.

Look at InputTrigger/AnalogSide for more details on solutions, particularly our favorite LM1815 chip.

See GenBoard/Manual/VrSensor/Polarity

Crank/cam trigger types and engine operation modes

4 (on an 8cyl engine) evenly-spaced crank-pulse can make injection and ignition work with a distributor.
a 24-1 or 36-1 or 36-2, 60-1, 60-2 or similar cranksensor enables wasted spark (no distributor).
a cam-position sensor for getting the exact phase is needed for non wasted-spark distributorless ignition or true sequential injection (with cranksensor if you want to time ignition precisely, since the camdrive is usually sloppy). This is not strictly true, since fully sequential inj and distributorless non-wasted spark can be done without a camsensor using IonSense, but anyone who can do it knows that he can :-)
For engines where a missing tooth trigger wheel isn't available I recommend using a crank trigger with 2 teeth for four cyinlder, three for six cyinlder and soo on. The Cylinder pulse should occur between 112 and 40 DegreesBTDC (but also at least 10Deg away from the closest TDC (makes the acceptable range 80-40degBTDC for V8)). This is combinded with a cam phase sensor close to TDC (at least 10deg from the closest cylinder pulse).

The final input-trigger answer is to have cam and cranktrigger as

well:

go fully DIS if cam and cranksignal are OK.
Go wasted spark if camsignal is lost (This will only work on an even fire engine! So no Harley, Ducati, many V6...)
Go fully DIS but retarted limphome if cranksignal is lost.

GenBoard has the IO and clocktics resources to handle both cam + crank inputtriggers at the same time. Handling multitooth trigger with the resources in GenBoard was not hard.

You can make a cheap timing light for your GenBoard:

Connect 5..7 high-power (narrow-lightcone) red LEDs in series, connect anode to +13.8V, and control catode with NPN and a 2 Ohm current-limiter resistor.

Also good for verification of your config :-)

Alien ign logging

Sorry, i'll make this nicer later..going home.

code must be compiled without LOG_ADVANCE_CHEAT in my_make, but with SEND_DWELL_NOT_IAC defined!

then configure secondary_trigger edge, enable & filtering

finally you can log the advance with megatun*. Dwell/advance are sent in the megatun* 'A' command (see comm.c)

Question: is a SYNC sensor needed to reset to cyl-1 after a full cycle?