While debuggin why I don't have spark, I tried to do a manual test of the coil & see if it can fire the plug and it doesn't
Here's my setup: I'm using Mitsubishi/Diamond F-608 coil. I connected one terminal on the primary side of the coil to +12V. Connect the secondary to a plug wire & plug. Ground the plug to the chassis ground. Then touch the chassis ground with the 2nd wire to the primary side of coil.
I get a good size spark from the wire but the plug never ignite.
hobieboy wrote:Then touch the chassis ground with the 2nd wire to the primary side of coil.
Okay - I'm a bit confused - you have a primary (+) and a Primary (-) and the Secondary HV spark plug terminal, yes? So it's +12v to the Primary (+) and just a momentary touch to ground to the Primary (-) to get a spark on the Secondary HV terminal. Is this what you're doing?
Yes - primary + is connected "permanently" to +12V. Secondary is connected to the spark plug. Then I touch the primary - momentarily to the chassis (which shows continuity with the ground).
To do this test you need an old style condensor/capacitor across your "points"
So +12v -> coil -> condensor ->ground
To fire the coil short out the condensor and release.
Without it you get a spark across your jumper wire instead of the plug wire
James,
How is this test differ than using VB921 to drive the coil then?
The VB921 essentially ground the coil (hence charge it). Then when MS "un-grounds" the coil, then it suppose to fire. And all these without the condensor?
hobieboy wrote:How is this test differ than using VB921 to drive the coil then?
The switching of a VB921 is very fast - which creates a stronger field in the coil than mechanical switching. It all comes down to how fast the current is switched on and off.
Points (or manual wire jabbing) are slow to switch due to arcing as they open. In mechanical ignition there is a condenser to absorb the voltage that causes the arcing to save the points, and although it therefore improves the spark, it slows the field collapse and overall secondary spark energy compared to fast and clean switching. The VB921 switches so fast it makes a stronger field and greater secondary voltage energy without the condenser that slowed the process before.
I successfully test with simple grounding , however I use a wire brush or file to create the make/break contact or at least a quick dragging action where some of the breaks are clean enough to get a spark. Most coils will fire with this method - albeit somewhat weakly. As per James' suggestion, if you're not getting a spark in your testing then adding a capacitor or condenser from coil (-) to ground may get you visible sparks. Adding a condenser to an electronic system should actually decrease the spark energy for the reasons above.
David, thanks for the clarification... What kind of capacitor/condensor do I need for this test then?
I am perturbed - I tested by VB921 circuit manually (simulating a 5V via a 680ohm resister into it) and it functions perfectly. MS is cycling thru the front LEDs so pressumably is switching the VB921 on-&-off. But I don't get any spark, hence these test.
I'd just use whatever condenser you have around from most any distributor or even from a lawnmower. Secondary spikes are large so a capacitor would have to be rated several hundred volts and a few uF I'd think.
Couldn't you test your output with a voltmeter? One probe to (+) source and other to VB output should see no circuit until it 'fires' (grounds) and then you should see a bit of voltage on the meter. That's how I'd try, anyway.
As you have tested the spark by manually operating the VB921 and it works then this confirms that the coil and VB921 is working ok.
A couple of thoughts:
a) It is possible that the LED driver transistor is placing a little too much load on the CPU pin and reducing the voltage that the VB921 sees. Try desoldering (or cutting) the "bottom" leads of R26, R27, R29 so that the CPU is feeding just the VB921.
b) by remote mounting the VB921 you are getting a voltage drop that prevents them from switching on. Try local mounting them or use a buffer/driver chip like the 74HC125.
To be sure, I haven't confirmed that the coil actually works. I manually switch the VB921 and it does give continuity (between collector & emitter) when applying 5V thru a 680ohm resister to the base.
But I thought about overloading the CPU as well. Maybe I'll try with a 74HC125 or similar & see - The 74HC125 triggers on low though, and I think the MS triggers the VB921 on high. Is there a tri-state buffer that triggers on high?
And to be sure - the circuit looks like: replace the input to VB921's with a trite state buffer trigger. Connect input of buffer to 5V, output to the 2 680ohm for the 2 VB921?
thanks.
EDIT - a 74HC126 is probably suitable here since it triggers high?
ok - tested it by removing all 3 LEDs and at least I get some sparks now
I say some because I have only visually checked cylinder 1.
Will add in the 74HC126 tomorrow when I get them.
Now, the BIG question - she still wouldn't start. I can hear combustion now.
And datalog shows that its oscillating between crank & ASE/WE.
So where do I go from here? I have tried a few different cranking angle and no luck yet.
And can someone please confirm/correct my understanding of using a bus driver for the VB921 -
I thought about tieing the input of the HC126 to VCC. Use the uP output to drive the enable (OE) line so that the VB921 either sees high or high impedence.
ok - did the mod and it is now causing MS to reset
I looked back at the previous logs - when there was NO ignition, I get very stable RPM. When I started to get some sparks (by removing the LEDs), MS started to reset a little and RPM is less stable.
When I tried the mod now, my RPM is like a yo-yo going between 0 & cranking range (~250). And MS is resetting quite a few times.
