SPARK ADV, SA-REF Blend Counter Increment per DRP

This value is added to the SA Blend Counter after each detected DRP to control SA-REF Blending when:
- SA-REF >= 0 AND
- NOT in DFCO
Factory Default = 16 Counts

Anyone know what this is/does?

Also, what is the "Blend Counter"?

 

I have not looked into this.

Is it possible that it’s a bit of logic that slows large changes in timing so as to not have violent swings? Basically a coef filter.

i assumed $8d had one but never found it nor spent much timing looking.

 

Hmmm....ive been a search and destroy mission for filter coefficient and lag filters lately....to a large degree of success in yielding much sharper and more consistent throttle response.

Wondering if i played around with this parameter. But I'm curious to know what it does first.

Ill try to get the address. But its in v7 of the SAUJP XDF.

 

Get the address and look it up in the anht disassembly.

There you can read notes to hopefully get a good idea of what it does.

 

good idea...

In the ANHT hac, it's under L802E which seems to correspond to the 0x2E in the xdf. I see it referenced in a couple of places...

16 BIN ADDED TO SA BLEND EACH DRP

16 CNTS ADDED TO SA BLEND EACH DRP

Not sure what CNTS or BIN means though in this case though...

 

Cnts is counts.
bin is typically binary but that’s odd for that there.

 

that much I figured... but what I'm not sure about is how that relates to actual commanded spark timing, or how the commanded spark timing is influenced by this.

And I know that a DRP is a distributor reference pulse

What exactly is it "blending"? The only other blending I'm familiar with in 8D is when it'll blend coolant and air temperature together to properly read the coolant sensor in the metal TPI plenum.

I guess going back to the XDF, there's something called a "SA-Ref Blender"

Maybe if I knew what that was, this might start to make sense?

 

Actually, there seems to be a related one...

Address: 0x5F3

XDF: DFCO, If NOT in DFCO, Counts added to SA-REF Blend per DRP

When in DFCO, SA-REF is set =0. Spark Blending is used to gradually return SA-REF to its calculated value by adding these Counts each time a DRP is detected.
Factory Default = 2 Counts per DRP

So in this case, coming off DFCO, returning the SA back normal?

 

Could be.

 

lol.... I plugged the question into one of these new-fangled AI interweb thingy's....

Here's what it said...

In the $8D mask (common to 1990–1992 Speed Density TPI systems), the relationship between "counts" and "degrees" is a matter of 8-bit conversion.
The ECM processes spark advance as a digital value between 0 and 255. To translate this into real-world degrees that you see in TunerPro or a data logger, a specific scaling factor is applied.




Common Reference Points
Because of this 8-bit limitation, you can determine the maximum possible advance and the resolution of the spark tables:
Counts (Decimal) Degrees of Advance Notes 0 0° No advance added to base timing. 1 0.35° The smallest increment the ECM can "step." 128 45.0° Mid-point of the byte. 255 89.65° Theoretical maximum (hardware limits usually cap this at ~42°).

Why 0.3515625?
This seemingly random number comes from dividing 90 degrees by 256 counts (


[img]data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACH5BAEAAAAALAAAAAABAAEAAAICRAEAOw==[/img]

In the GM ignition logic, the distributor reference pulses occur every 90° of crankshaft rotation (for an 8-cylinder engine). The ECM uses this 90° window as its mathematical "sandbox" for calculating when to fire the next spark event.

Key Limitations in $8D

• The "42-Degree" Ceiling: Even though the math allows for 89°, the physical design of the small-cap HEI distributor usually limits you to roughly 42° BTDC. If you command more, the spark may "arc" to the wrong terminal inside the cap (cross-firing).
• Initial/Base Timing: Remember that the ECM doesn't know where the distributor is physically turned. If your base timing is set to 6° at the balancer, but your BIN file says the Reference Angle is 0°, the engine will actually be running 6° more advance than the counts/formula suggest.

 

So what ever this blend thing is doing from my first post, it's adding 16 x .315625 = 5.6° of advance back in per DRP. If the SA gets set to 0 for some reason, it'd take ~8 DRPs to get back up to my typical spark advance of ~40 (for part throttle situations).

Or coming out of DFCO, it's adding 2 x .315625 = .63° of spark advance back in per DRP.

 

So I guess other than knock retard, when would the ECM be reducing the SA-Ref? Such that timing has to be "blended' back in with this function?

The knock retard has its own recovery rate table, so it doesn't seem that it'd be using this blend function??

 

I was looking to add such a thing for gear shifts to prevent shift knock. Which the ecu does not do.