How do Atom Computer Ignition modules work?

[legendre]

Quick update.

Good friend loaned me an early 1950s B&S model 5S motor for a test-bed. From what I understand this engine is rated for 1.0 HP @ 3200 RPM. Bore 2.00" / stroke 1.5" / disp. 77cc. Compression seems ok, but haven’t measured it yet. It starts easily and will idle reliably – but it is burning some oil and the plug has coke built up in the cavity. Seems it was the power plant for a mini-bike back when, and it’s definitely well-used.

Best part, is that it started & ran without trouble using the breadboard version of the BU931 based module. Haven’t had a chance to measure anything yet, all I’ve learned is that the module is definitely phase / polarity sensitive on this motor. Hope to get some timing & RPM measurements later today.

[phil-b]

quote legendre:

quote debe:

Do you have any sense as to why the Mosorb raises the upper speed limit (essentially, gives you more HF bandwidth)? At the moment, I can’t think of any possible mechanism, other than some device-specific quirk.

Electronics can do the craziest things.

I’m a software guy, with a general black thumb with hardware, but….. I suspect the nasty transients happen AFTER the spark, and are of higher voltage than seen before the power transistor is turned off (points open).

I was surprised that in your case the spark came back, ever, after it stopped!

[legendre]

I had a chance to run the engine this evening, and take some measurements. The results are pretty surprising.

Setting up:

Determining precise TDC on my test motor is difficult. Due to the plug hole being offset from the piston crown, there is no straightforward way to determine TDC using either a piston stop or a dial gauge. I had to eyeball piston height through the plug hole – made several observations and took the average of them.. but I think I’m within +/- 3′. Made a datum mark on the case with a file, and a TDC mark on the flywheel with a Sharpie pen.

Likewise, determining the fire point was a problem. All I could do was to set the points per the manual at 0.020" as carefully as possible, and then mark the ‘fire’ point on the flywheel with a sharpie using a beep-box to determine when the points broke open.

Measuring:

Started the motor, and used a strobe timing light to view the timing angle. Oddly enough, at idle and any low-ish RPM, the motor (running off the module, now) was firing +nearly spot-on+ with the ‘fire’ mark I’d previously made. That’s just got to be a coincidence, I’d think. But I repeated the start-up & measurement four times, and each time, the dang thing is hitting within +/-2′ of the ‘fire’ mark. So it is!

But here’s what’s more interesting. As the RPM is increased, the timing +retards+, rather than advances.. as one would expect. The retard isn’t highly significant, maybe 5-7′ tops, at full throttle. But it does leave a number of questions – such as, why doesn’t it +advance+ with speed, as one would expect from any other ignition system?

Crazy stuff! Any ideas, gents??

[Mumbles]

The module doesn’t know how to advance the timing. It wasn’t designed to. The original Red Atom for outboards work the same but spark advance is achieved by rotating the mag plate. Briggs motors have no mechanical advance system.

[debe]

Try using the .47uf capacitor, instead of the 1uf as capacitor size does alter things.

[legendre]

quote debe:

Try using the .47uf capacitor, instead of the 1uf as capacitor size does alter things.

Probably so.

I need to order-up some low-ish voltage caps for this use – 0.1uF, 0.22uF, 0.47uF, 1.0uF etc. The 1.0uF electrolytic I’m using in my module is by no means optimized. All I know for sure is that is +will+ run the motor, and the timing doesn’t shift wildly at RPM.

Let me see what I can do about some appropriate caps from the eBay.

[samson]

I just wanted thank debe and everyone else that contributed to this thread. I made a pair of these modules for my 1974 Evinrude 15 hp outboard that has a worn out timing plate. I tried a pair of Nova IIs first but they just fry after about 2 minutes of run time. My motor is an electric start version and has 4 magnets in the flywheel so it fires each cylinder twice per revolution which is I why I think the novas didn’t last very long. Or maybe cause of voltage spikes like debe was suggesting but I’m not as savy him on coils. The modules give a nice bright spark and lasted the duration of the 10 minute test that I put them through. This motor revs to between 5,500 and 6,000 rpm and they didn’t miss a beat at WOT. I used the 2SD1071 found here : http://www.ebay.com/itm/121097158690 and the 15ke200A for the diode. I started with very little knowledge of circuits and have only made one similar item in the past, a momentary pulse relay for a pig trap door actuator.

Yall don’t make fun of this Texan’s soldering skills! Scary that these look better than my relay!

[mercuryman]

You will need to mount the power transistor on a heat sink and insulate the heat sink from transistor. Looks good to me.

[debe]

First point , you have done a good job & you will get better with practice & it is a fiddly job to do. Next point is for reliability don’t use paper type electrolytics as in your picture, they will dry out & fail in a hot environment. Use Polyester MKT types as in the original pictures. I have used a small piece of flat aluminium for a heat sink, insulating the transistor from it. Make sure none of the wiring can short betwean each other, then seal it all with epoxy resin or I have used hot glue gun. This then makes it weather proof.

[legendre]

quote Mumbles:

The module doesn’t know how to advance the timing. It wasn’t designed to. The original Red Atom for outboards work the same but spark advance is achieved by rotating the mag plate.

Ideally, the Red module shouldn’t create any dynamic advance or retard – following only the advance curve of the movable stator plate. But per my understanding of the laws of induction, a circuit of this type should +tend+ to produce an advance behavior with increasing RPM, unless the circuit is additionally designed to squelch it.

quote :

Briggs motors have no mechanical advance system.

Far as I know, no they don’t – at least not the typical lawnmower / snowblower motor we’d be most familiar with. Doesn’t mean it might not be advantageous, though! Sure, the motors run ‘right’ with fixed timing, but if cost / complexity were not an issue, I’d expect to see dynamic advance on these motors.

quote debe:

(…) Next point is for reliability don’t use paper type electrolytics as in your picture, they will dry out & fail in a hot environment. Use Polyester MKT types as in the original pictures. (…)

This is a very Good Point. I’ve been using aluminum electrolytics in my modules, simply because I had nothing else on-hand in that low value. I know it’s not helping the performance of my modules, and plan to get rid of them ASAP.

The AEs may even be responsible for the retard-at-speed behavior I’ve clearly observed.

quote mercuryman:

You will need to mount the power transistor on a heat sink and insulate the heat sink from transistor. Looks good to me.

FWIW, the (Darlington) transistors in my modules have never gotten even warm.. this with the either of the successful devices.

Were the original modules heatsinked? I don’t think so.. or at least, not robustly.

Looks like they have a little brass ferrule that admits a mounting screw though the hole in the TO-220 case. But none of this provides the intimate contact you’d expect in a heatsinked part.

The plastic case on the Atom module is a poor thermal conductor.

[Author]

Posts