Inliners International Forums General Discussion Engines spark lead, angle vs. time

I just stuffed a ford EDIS wasted-spark system in my inline 6. I basically cloned the distributor "map" (mechanical X vacuum) and tweaked it up a bit, seat-of-the-pants, and it's working GREAT.

Now I want to do it right. I don't have access to a road dyno, etc.

I know that spark lead is usually expressed as an angle (eg. distributor rotation) but it's really a lead TIME for the flame front to propagate. For argument's sake, assuming that F/A ratio, cylinder filling, etc, is constant across some RPM range, the question is: should I treat spark lead as time, in milliseconds, and for each point in my map, convert that lead time to degrees?

If that tends to hold true, then I have a basis for filling out the map. For RPM 1000 to 5000 at WOT, assuming my carb is perfect (!RIGHT!) spark lead might be some constant number of milliseconds, but more degrees of lead as RPM increases.

A secondary question is just how the hell do I determine optimum spark lead for a given rpm/load. I'm thinking that Smokey Yunick might have used a pyrometer to measure exhaust gas temp, and tune for lowest exhaust gas temp (both mixture and spark).

is this crazy talk?

Hi Tom . . .

Here's a really good conceptual write up on ignition timing from the superhonda.com website.

The author lays out some of the basic arithmetic involved in finding the sweet spot. I think it does a great job of helping one visualize the impact of to much or too little spark lead . . .

regards,
stock49

Hey thanks for that! It's as i thought, but his clear understanding of the math is great. That's a great explanation of peak pressure and optimum ATDC angle.

I made a spreadsheet of his calcs. There are two variables I don't know: spark gap to piston crown distance, and flame propagation speed. It reproduces his result.

I can't find a record of my piston-to-head clearance. Stock pistons rods crank and gasket, I probably didn't measure it.

Flame speed: umm this ain't no Honda, so I'll assume it's the low end of the range.

Assuming piston/head clearance of .06 (8:1, complicated wedge with steps and clearances) and optimum crank angle of 13 degrees:

RPM timing from TDC (- is BTDC)
2000 0.8
2500 -2.2
3000 -5.3
3500 -8.3
4000 -11.4
4500 -14.4
5000 -17.5

Factory spark timing is 16BTDC static+mechanical around 2250, which would be way too much. If I set flame speed to 700 meters/sec then it's correct around 2500, but the slope is wrong.

This credible page
http://www.autometricsmotorsports.com/techtips/02.htm
says gasoline flame propagation is 15 meters/sec, not 1.35-2.2

http://www.faqs.org/faqs/autos/gasoline-faq/part3/section-1.html
says 16.5 m/S (for 12:1)

So I don't know what to think.

(3.125" bore x 4.25" stroke, 6.85" rod length)

The ignition process begins during compression stroke. Increased pressure and temperature causes precursor chemical reactions. Then spark ignites the flame front which takes some time as well to grow from the actual spark to a flame front. Then the flame progresses across the chamber. The last part of the mixture to burn is the end gas which is highly compressed with plenty of time for precursor reactions so it can spontaneously ignite also known as spark knock.

All of the above process is sensitive to engine speed. Most notably the flame front progression and end gas mixing at higher engine speed causes the combustion to occur in less time.

Note that ignition lead mechanical advance curve typically slopes up to a maximum at, say 2500rpm, then does not advance further. This fixed lead above 2500 is because the total time for combustion is less at higher speed roughly in proportion to engine speed.

For determining what ignition lead you need try this: Run some crappy low octane gas like 87 no ethanol if you can find it. Play with your timing gradually creeping up on your map. If you hear some spark knock at certain points then adjust accordingly. Repeat until you are happy. Then put some better gas in it and drive on.

Hi Tom . . .

Unfortunately, there are a lot of variables here. From what I have read flame speeds are dependent on fuel, additives, stoichiometry and turbulence or lack thereof (the so called Reynold's number). And I don't think the leap from theory to practice is all that straight forward.

Google books has got a very scholarly work online (Charles Taylor):
Internal Combustion Engines in Theory and Practice

This author expresses flame speed in ratio to piston speed - his graph suggests that flame speed is not a constant in these calculations. (echoing strokersix's post)

regards,
stock49

what's funny about all this, is i get grief from people (not inliners, i'd like to point out! :-) about "advance til ping, back off". i've tried multiple times to put some rigor into tuning.

the supertuner website info is correct, as far as it goes -- the problem is that the dominant variable is virtually unknowable -- flame propagation at at particular load/rev/chamber shape/stoichiometry point.

i'm very glad i made the spreadsheet though -- it allowed me to do what-ifs. Often, this sort of process in other areas (gearing, cams even, chassis stuff, electronics, ...) converges on a workable range of values from which to choose and tune. In this case both the slope and intercept of the spark curve (assuming a linear approximation) is all over the place.

so basically i can now justify what we've all known and done -- start with what you've got and tune on the road!

I still want to have some physical justification/verification for settings though. I will pursue exhaust gas temperature, first to get F/A mixture right then for spark.


However, i had NOT done the math on my engine until now -- the loooooong rod and stroke and slow turning mean that I really don't need a lot of spark lead -- unless the combustion chamber shape, which really is bizarre and terrible, hugely slows down flame propagation.

Thanks for humoring me here.

Thanks for the google book link, i'll spend some time with that this week.

"the problem is that the dominant variable is virtually unknowable"

This is why I think the poster from the superhonda site chose to make some assumptions to simplify the math and drive toward conceptual understanding . . .

I think you are spot on with the implications of long stroke and low RPM. Page back to Figure 1-12 in Taylor's book. I think such engines have a more shallow 'effective burn angle' and as such can tolerate only so much spark lead.

The one thing that I have always had a hard time getting my head around is the presence of auto-ignition versus 'ping'. I don't think that they necessarily go hand in hand. The idea that 'ideal tune' is where the flame-front smoothly consumes the entire charge just as it reaches the piston head seems overly theoretical to me.

In practice I think that smooth combustion consumes MOST of the charge - with the rest auto-igniting. But since auto-ignition pressures are occurring well after TDC (and for just a small percentage of the charge) there is no 'ping' associated with them . . .

Perhaps this is why advance-till-it-pings-under-load is a persistent rule of thumb . . .

It is possible to have too much ignition lead without spark knock. This is why I suggest using crappy fuel for test, to encourage knock for tuning purpose. If you try to tune "advance till ping then back off" with good fuel you may end up with too much ignition lead. Result will be increased fuel consumption and overheating.

The danger from too much spark lead is not the pinging you hear when pulling a grade at part throttle.It's the detonation occurring at wide open throttle when it's hard to hear it over the engine roar.Repeated full throttle detonation breaks pistons,cracks bearing caps,causes crankshaft breakage and of course reduces power.This destruction won't happen from an occasional 10 second full throttle blast but doing it longer and more often can be an issue.
True proper total spark advance is most commonly done on dynometers. Then the basic settings can be transferred to other engines and adjusted slightly for conditions.It can also be done with repeated timed acceleration runs,the drag strip.
I don't believe any simulated situation can give you accurate timing for a real engine,close maybe.
Who here knows the best spark advance curve and maximum degrees for a 235 or 261 Chevy? How about 270 GMC?

I think I'd open another cold one an'try not to think about it... that an' pumping losses.

There's a somewhat vague method of deconstructing observed behavior to guess at WOT spark. It's based on the idea that the chamber function will be similar under changing load conditions.
Attach vacuum advance and run with existing curve, motor at full temperature, normal plug heat range, etc.
Monitor vacuum gauge for highest vacuum at 50 mph cruise. Advance until the reading stops going up (i.e., no change), and retard to previous setting. E.g.:
40° = 16" vacuum
42° = 17" vacuum
44° = 18" vacuum
46° = 19" vacuum
48° = 19" vacuum
Use 46°
2/3 of this (30°) will be pretty close for WOT spark. How much is initial vs. mechanical depends on chassis weight, gearing blah.

In the old Jr. Stock drag racing days I ran a 57 Chevy Sedan Del. in K/S Auto. in the 13.50 et range and I use to do two things when shifting to Hi Gear at about 1/8 mile and that was with a hand control I would dump most of the transmission pressure and at 6000 rpms with a cable attached to the dist. back off about 10 degs. of lead. Doing this was good for about a tenth and 4 to 5 mph in the 1/4 mile.

You may be onto something with the cold-one preacher edski - might provide a quench effect . . . as for pumping losses - no worries - we're talking WOT so there aren't any . . . :-)

Thanks for the suggestions, opinions, ideas, and beer. I think I need one too.

While I was playing with my distributor i realized hmmm i've never heard this engine ping? so i kept advancing it... advancing it... (i know the dangers of detonation so i was pushing it in small controlled bursts) .. finally it just plain acted weird. it did make a noise, but not what i'd call the marble in a beer can "ping". too much cast iron?

so it's hard to gauge by ear, part of the reason i'm looking for a metric. i will try the vacuum/50mph cruise technique -- now that i can set timing from the drivers seat (no more dangerous than texting :-) That method will let me at least sanity-check a cell or two.

my phone has an accelerometer (what's this world coming to) and i was thinking of trying that out to see if i can get consistent results.

i also pharted around with the spreadsheet based upon superhonda's data, tried to fudge between likely bounds, it's very little advance across the board, i will 'spearmint with that.

i have a nice freeway drive to work tomorrow so that's the test bed!

thanks for the hints and help, if i blow it up, i'll send you all the bill :-)

Grumpy Jenkins used to race a (I believe) a '56 sedan delivery too and his high gear timing retard trick was to use the dual point 'vette dist and separate the two points and run on the advanced point set, until in high gear then flip a switch to run off of the retarded set (approx 9-10 degrees diff).

My son an I bought an Accel dual point conversion kit for his 462 (.030 over 455) Pontiac that was having a time trying to start against 10 degree static timing during the 100 degree Texas summers. We wired one set to the solenoid (retard set) so it would only be in use while the starter was being spun and after releasing the the starter and the solenoid not being used, Ignition duties were then provided by the advanced set. The big poncho would really spin over with that retarded timing and run stout and snappy with the 10 degree of so static timing!! kinda like having your beer and and cake too(kinda of a Bohunk thing-I guess).

stock 49--
...an' heres to quench... burp!

OK, back from the trenches [OK, LA freeway] reporting back on the results of actually calculating spark timing, using the superhonda formula. In short: it's a crock of poo.

The problem is he's got the single most crucial factor trivially set to a flame-speed constant times the combustion chamber *height* around TDC. It's completely and utterly wrong.

The mechanical calcs are probably OK for a first approximation; it more or less assumes that the chamber volume is constant during the flame time. (we know this is not true).

The gist of it is this: for all speeds the charge burn time is assumed to be constant, and very brief -- flame speed multiplied by the combustion chamber *height*. That's nonsense; the chamber is only .030 - .050 whatever high, but it's inches in diameter, duh, the flame spreads *sideways* and all-ways. Plus it's dynamic -- don't need no theory to notice spark lead isn't simply proportional to RPM.


According to his calcs -- with anything in his recommended flame speed range -- my spark lead should be between 0 and 10 degrees. I know from careful experimentation that it wants more or less 26@2500 WOT, etc.


So -- assuming that 15 degrees ATDC is where max pressure wants to be (close enough for this purpose) -- and going from OBSERVED REALITY, I back-calculate that my effective charge burn time is about 13 milliseconds at 500 rpm (10 degrees lead), and 8.8 mS at 2500 rpm (26 degrees lead) at WOT. That's more than 20 times off his simplistic calculation.

OK my mixture varies with air speed (carb) where FI is probably more linear. Still, his calcs are useless. The point is, it's far too complicated to have a one-size-fit-all formula.


The answer is what we always knew: start with something (like stock) and tune til correct.

Does anyone know the typical stock total advance in degrees used on a 235-261? I do know a stock GMC 302 distributor has about 26-28 crankshaft degrees of mechanical advance plus the 5 initial.
It's my understanding that drag racing engines often run extreme ignition advance through through the lower gears.
I used to race British Triumph 650 motorcycles.Total advance was around 42 degrees on a 20 cubic inch true hemi head cylinder turning 7500 rpm.
Most all wedge head pushrod V-8's run 34-40 degrees total advance for street engines.Newer SBC quick burn head designs about 30-32 degrees.
I'm thinking flatheads use less advance?