HEI with Electronic Spark Control

Good work - I'm sure it's gratifying to be the first on a useful hack/mod like this.

So, you're running the ESC module in open loop mode? From your previous post, I thought that required an external module of some kind?

 

Stage 1: I'm currently in what could be considered a "limp home" mode. By disconnecting the harness, there is an open line and the HEI assumes module is not working. The system is designed so that, in the event of this module failure, the distributor adds a small amount of advance when it goes above 800 or 900 RPM and just stays there. It's kinda neat watching it on the timing light; unlike vacuum and mechanical advances that are very analog and linear in operation, the ESC runs rock steady at initial advance, then instantaneously steps the timing up when it reaches it's programmed RPM.

Stage 2: I'm going to try to come up with some type of simple circuit that tricks the HEI into thinking it's received a full advance curve. When I say simple, I mean possibly just a single diode. Just going to have to see what kind of output is on the line going to the module; everything I read says it's just a ~5vdc square wave pulse that is converted from the AC of the pickup coil.

Stage 3: Try to figure out how to get an accurate knock sensor reading on the Dauntless and run the system off of a module.

Stage 4: Do the same project with a small cap HEI.

 

Scope time.

Supposedly there is some voodoo to this. You can try mounting the knock sensor on the block in the same manner as the original engine, and see if you get a signal. You'd have to induce preignition somehow I think ... not sure how you'd do that. I'd look at a 4.3L to see where the KS is mounted, and try to duplicate that... this assumes that the acoustic/mechanical coupling will be the same for the two engines. Clearly not, but maybe close enough.

Fun project. I'm glad you're making progress.

 

Well, doing some initial leg work on this, realizing that getting the dist to run a full advance curve standalone is going to be a little harder than I thought unless I want to spring for some go-fast parts ($$$) from MSD or Mallory. Thinking about trying this little experiment first:

What if I used a 555 or 556 IC? You can get them for a couple bucks. I think a true advance circuit would be hard to do and creeping outside the scope of this project (why re-invent an ECM?), but what if I could build a monostable delay circuit at my kitchen table that would fit into the base of the HEI next to the ignition module? Thinking out loud here, but let's say I want around 32-35 deg total advance. What if I could build a circuit that would delay 20deg or so at idle and zero at 2500 RPM? Then, you would just set initial advance to 35 timing. When you started the engine, 35deg initial - 20deg digital delay equals 12deg total timing. As RPM increases, delay drops off until 2500 RPM where delay = zero. At 2500RPM, 35deg initial - 0deg digital delay = 35deg total timing. Sounds plausible to me, but I'm more of a circuit fixer guy than a circuit designer. I would use the 5vdc square wave from the ignition module as a trigger.

Any thoughts on this?

Does anyone know of a cheap, off the shelf module that would already do this for me?

 

Also thinking aloud . . .

This sounds like a fairly easy project, assuming a few things. First, Is the advance "curve" linear? Is it dependent only on RPM, or are there other factors to consider?

Of course, the 555 operates in dT rather than degrees, and you might have to settle for something more than 0 delay. How many microseconds is 20 degrees at idle? How many microseconds is 1 degree at 2500 RPM? Given these endpoints, it should not be hard to work out the circuit. Maybe just get it close and fine tune it with a pot.

 

'Zactly what I was thinking, too.

 

I started to do some calculations and realized I don't know what the advance angle relates to. Does 35° mean 35° rotation of the crankshaft, 35° rotation of the distributor shaft, or 35/360 of the time between points openings?

 

Timing is in degrees of crankshaft rotation, ie 5 degrees before top dead center. Thus you rotate the distributor 5 degrees for every 10 degrees of advance.

Regarding your square wave generator, I'm sure there has to be a waveform generator on a chip. It'd be similar to a 555, but likely fewer external components, and more stable. I'd wonder about thermal drift with a 555 ... useful device, but quite rudimentary. Searching... easy search: http://www.siliconfareast.com/waveform-gen.htm - take a look at the spec sheets and any app notes for these devices (ie http://www.intersil.com/cda/deviceinfo/0,1477,ICL8038,0.html ). You should be able to source something like this easily through Mouser or Digikey. I would certainly build a regulated supply for it, on the same board, and add some kind of transient protection (maybe as simple as MOVs or zeners across the supply rails).

 

35 degrees crankshaft rotation, converted into milliseconds.

Tim, I was thinking I'd just draw the 5vdc square wave off the 7pin HEI module (EST pin), delay it, and output it back to the module on the input pin, I forget what it's labled at the moment.

 

You should be able to make a programmable delay with a 74LS122 MSMV or similar - a 50c part. Use the first MSMV to control the duration of the delay. Trigger the second MSMV on the falling edge of the pulse from the first MSMV to make your output. 74122 is a dual package IIRC.

http://focus.ti.com/lit/ds/symlink/sn74122.pdf

This should be fine if you only need a constant delay. No guarantee, since I have not breadboarded this circuit, but I think it will work.

Edit - no this won't work. Delta t gets smaller for the same degrees of advance as the engine speeds up. You need a circuit that responds to the engine speed to change the delay.

 

an LM1815 would let me run straight off the pickup What about a 555 divider; would the oddfire pulse not be a possibility?
Posted via Mobile Device

 

I don't know what's going on in the oddfire pickup coil, but I guess you could make your own detector with that and a LM1815. I don't see what the advantage is though - you already have a pulse that is normally sent to the ECU. The GM ECU can't handle this signal because of the odd interval between the pulses, AFAIK.

What input does the module expect for advance? Is it synchronized with the module firing or does it just inform the module how much the advance should be? Really, all the module needs is some kind of signal that is proportional to the advance setting. The module is capable of calculating the advance, as evidenced by the function of the limp-home mode. So, the signal needed could be very simple, ie a PWM input that is proportional to the advance that the ECU is calling for. Or, on the other side of the coin, the ECU could be calculating the timing verbatim, and sending a pulse whose leading or trailing edge actually triggers the spark.

I think the former is more likely, but that is just speculation. If it is a PWM circuit, then all you have to do is send back a pulse length triggered by the ECU input that is the right length for full advance.

 

So I've had EFI on the brain for a couple days now. Been reading about various manufacturer's ECMs from various different eras. Head is starting to hurt. Anyway, they all seem to have the same basic concept of operation. They use tables of values to compute various different TBI outputs relative to sensor inputs. Some of the tables relate to spark advance, and on the GM '747, one table in particular is the spark advance/retard value based on other sets of computations.

I was thinking about how I'm going to control spark on an odd fire engine using an even fire ECM. I know that the HEI module itself will cope with the odd fire reluctor; my HEEP is running proof of that. What I'm not sure about is how I'm going to get the ECM to predict an ignition even that won't physically occur at the time when the odd fire crank will reach it. The problem is with the advance timing, in particular. Specifically, the calculations for it are made from the trailing edges of the HEI module's pulse that preceeds it. Obviously, that won't work with an odd fire because of the uneven firing sequence. Well, what if we just do what we always do when we need to move timing around on the odd fire ignitions: Move 2 spots at a time in order to keep your firing sequence in phase with the reluctor.

The ECM has an advance table. Say it wants to advance 30 deg total. 6 deg BTDC is built into the base. 24 deg is calculated and done electronically. There's no servo like there is with a vacuum advance module. But what it's actually doing is delaying that spark 120 deg (crank rotation) minus some value of advance until the next piston comes up, then fires. Well, what if we just delayed it an extra amount? In fact, what if we could change the table to delay every single pulse, in serial, coming from the HEI an extra 240 deg (crank rotation) to come up on the next pair of firing sequence? Afterall, an odd fire V6 is really just 3 V-twins working in series.

#1 pulse would be used to calculate advance, then delayed 240 deg to fire piston #5.
#6 pulse would be used to calculate advance, then delayed 240 deg to fire piston #4.
#5 pulse would be used to calculate advance, then delayed 240 deg to fire piston #3.
#4 pulse...
#3 and so on...
#2 and so on...
#1 repeats the sequence.

Will be a couple months before I have time to 'scope this theory.

Thoughts?

Anyone with GM '747 table knowledge know if 240 deg (crank) delay is within ECM possibility? Basically just wanting to double the amount of time the ECM delays the pulses going out to the HEI.