How do Atom Computer Ignition modules work?

[debe]

Post some good clear pictures of what you have built, otherwise were just guesing. Mercuryman did you use a new BC337 when replacing the wrong connected one? They are cheap enough.

[mercuryman]

NO, I have a lot of parts I am going to try again.

[legendre]

@debe

Do you have any suggestions on subbing a part for the BC337? I don’t have any BC337 but I do have other similar devices, like 2N2222, 2N3904, NTE123, etc.

Have you tried any other types in that location? As mentioned, my circuit +is+ apparently triggering, but I’d rather duplicate your configuration closely as possible at first, and then strike out and test modifications once I have a known-working system. If one device I mention is better than another, I’d like to know.

ETA: Went and ordered up some BC337, but would still like to know if there are any good subs in the meantime.

[phil-b]

One reason I wanted to see if I could get PCB (printed circuit boards) made was to make it easy to experiment with different components (tho debe did extensive work in this department finding the best power transistor to use)…

[legendre]

quote Phil B:

One reason I wanted to see if I could get PCB (printed circuit boards) made was to make it easy to experiment with different components (tho debe did extensive work in this department finding the best power transistor to use)…

Unless the PCBs were quite cheap, how would that make experimentation any easier? PCB traces & holes can swiftly fail and peel / pull off.. Not so with simple perf-board and wire connections, which can be reworked again and again.

To me, a strip board aka Veroboard build would seem ideal, unless the target is mass production. If only I had the skills to lay it out.. I’m just awful at that!!

[phil-b]

It seemed like there are fabs that are a dollar and under per square inch for two sided (in China):
http://pcbshopper.com/

A dollar a card would be cheap enough for me.

[debe]

The reason ive used a BC337 is that’s what is in the original Atom module. Its an NPN transistor Vce (collector-emitter voltage) is 50V, Ic (DC current) is 800Ma. Also the pin out conections are diferent on some other types of transistors. I have kept to the original component values as used in the Red Atom module. The only part I had to find was a suitable switching transistor that would work in the circuit. Even that in some magnetos needed an extra protection device, particularly with Wico magnetos in Seagull outboards. BC337 transistors are cheep here in Australia on Ebay localy for $7.00 for 50 freight free. Ken W don’t quite know how you got away with using the BD649 for so long. What was you using the module on? The killer on these systems is the HV spike as the transistor fires the coil, and its diferent on just about every brand of coil. The old Villiers coil was not as harsh as the Wico coil on the modules. When testing with the Wico magneto eventualy it would stop sparking, if left for a while it would be ok again. By fitting the Transorb diode across the switchlng transistor it cured the problem.

[ken-w]

So far I have gotten away with the BD649, is because I have only done hand spin testing with an OMC universal magneto mounted on a gale 3D15. The piston is stuck in the cylinder so I have removed to get it out. Issues might come up when I mount this module into a running outboard. I did get the 2SD1071’s in and have tried to get them to fire. No such luck. Could it be that the emitter/base voltage is higher on the 1071(6v) then the 649 (5v)?

[legendre]

@debe

No worries, I found a really inexpensive source for BC337 on eBay. USD $3.20 for 50 pcs. including post, and that’s from a USA seller.

Not sure why, but those BC series parts and the like are not as common in the US as they seem to be in western Europe, the UK and Commonwealth nations.

[legendre]

@All

I finally had some time to devote to understanding this system, and I believe I’ve figured out how the circuit is supposed to work.. correct me if I’m wrong on this, as I’m not much of a solid-state guy.

As a cycle begins, the module is "off" and presents a virtual open-circuit to the primary of the mag coil. Once the voltage across the coil rises to the level required to turn-on Q1 (which might be somewhere between 700mV and 1200mV being a Darlington device) current flows through R1 into the base of Q1 and Q1 turns ‘on’. Q1 has huge current gain, and it turns on very hard, which now presents a near short-circuit across the mag coil. The mag coil current now flows hard through Q1.

The voltage at the R2-R3 junction is too low to do anything with Q2, yet – it’s only 25% of the voltage across Q1, so Q2 remains ‘off’. As the cycle continues, current in the coil builds as does the voltage across the coil. Eventually, the voltage at the R2-R3 junction rises high enough (about 700mV) to turn-on Q2. When Q2 switches ‘on’ it shorts R1 to ground, starving the base of Q1 of current, causing Q1 to abruptly switch ‘off’. C1 has the dual effect of delaying the time at which Q2 turns ‘on’ and ensuring that Q2 remains on long enough to permit a full discharge of the coil.

When Q1 switches ‘off’, this has the same effect as opening the points – the coil current breaks, the magnetic field collapses, and a high-voltage is induced in the mag coil secondary, firing the plug. Just as with points, a second, lower voltage (but still several hundred volts) also appears across the mag coil primary – but the transorb diode takes care of that.

So.. changing the ratios of R2-R3 should alter the timing, as should changing the value of C1. Either should also effect spark quality (voltage and duration). As for R1, I’m not sure how critical its value is – it seems to be primarily there to limit the base current of Q1, lest it become damaged. it would seem that R1 should be the lowest possible value that keeps the base current within the design limits?

[Author]

Posts