Overheat

[Indycars]

If I really stomp it I can see 5# for a few seconds, then I'm going waaay too fast!

Go to Bonneville, you will have lots of room to get it straightened out if if get sideways!

 

[PotvinGuy]

I have a question. HOW DO IT KNOW?

ORF, I guess you are asking re my advance gauge. It detects TDC and cylinder firing and computes the advance.

 

[PotvinGuy]

Go to Bonneville, you will have lots of room to get it straightened out if if get sideways!

Sadly, I lack the personal moxie needed to go much over 60!

 

[Indycars]

Sadly, I lack the personal moxie needed to go much over 60!

I was hoping you would say something like that .....
Ok, so you be my sponsor and I will drive. That way all you have to do is scream like this kid!!!



Or Suck it up like this baby!
.

 

[old round fart]

I guess what I was asking was how are the sensors configured to get data.

 

[PotvinGuy]

From my original design paper:

The firing signal will be taken from the coil or distributor. I'll assemble the ignition system on the bench and tweak the conditioning circuitry to get a clean trigger.
The TDC signal comes from an infrared sensor pointed at 4 reflective markers on the flywheel, spaced 90 degrees apart. The relatively large diameter of the flywheel offers more precision than any other part of the rotating assembly, and it hides the pickup and wiring, desirable in my application (exposed hot rod engine).

 

[Island Girl]

From my original design paper:

The firing signal will be taken from the coil or distributor. I'll assemble the ignition system on the bench and tweak the conditioning circuitry to get a clean trigger.
The TDC signal comes from an infrared sensor pointed at 4 reflective markers on the flywheel, spaced 90 degrees apart. The relatively large diameter of the flywheel offers more precision than any other part of the rotating assembly, and it hides the pickup and wiring, desirable in my application (exposed hot rod engine).

Kind of guessing at how you plan to work this . . . . essentially, you're triggering a counting circuit with the "coil fires" input and based on known, (or already measured) RPM, you're able to convert the count, which represents the interval between the "coil fires" input and TDC input, which would stop the count, into degrees . . . . ?

The higher the count, the more degrees before TDC, for any given RPM

I suppose making the count rate inversely proportional to RPM, would automatically account for the RPM in the actual count, and integrating the count pulses would allow for an analog meter type of readout.

 

[PotvinGuy]

IG, I don't usually reveal my secrets (the SEMA rats are notorious for stealing ideas), but I do have a Patent Pending on this and I get my jollies from suing big companies, so here is the whole paper. I haven't updated it at all (shiny objects distract me), but the concept is still the same. This link should work: TIMING DISPLAY.doc

 

[Indycars]

That is so cool, thanks for sharing!

 

[old round fart]

Wow! That is incredible. Hat is off to your you sir.

 

[Island Girl]

IG, I don't usually reveal my secrets (the SEMA rats are notorious for stealing ideas), but I do have a Patent Pending on this and I get my jollies from suing big companies, so here is the whole paper. I haven't updated it at all (shiny objects distract me), but the concept is still the same. This link should work: TIMING DISPLAY.doc

I have to look at it again when I get home from work, but I don't see that you have accounted for RPM.

Unless there's more to it, the pulse width, which is what you're integrating on the meter, will be the same length for a low rpm with little advance as with higher rpm with more advance.

 

[Island Girl]

I have to look at it again when I get home from work, but I don't see that you have accounted for RPM.

Unless there's more to it, the pulse width, which is what you're integrating on the meter, will be the same length for a low rpm with little advance as with higher rpm with more advance.

Looking at it comfortably from home . . . .

Let's plug in some realistic numbers:

Looking at 600rpm;

10 revolutions per second

Coil fires 4 times per revolution, or every 90*, or 40 times per second (however you want to look at it)

Time between subsequent coil fires, or 90* increments of rotation, is 1000ms/40 or 25 milliseconds (ms)

At 600 rpm, rotational speed is 3.6* per ms

At 1200 rpm it would be 7.2*/ms

At 2400 rpm - 14.4*/ms

At 3600 rpm - 21.6*/ms

With a speed of 600 rpm, assume you have an actual timing of 23.6*, so the coil fires then and the pulse lasts 1ms until it's killed by the trigger from the 20* infrared reflector sensor

Now let's look at a speed 2400 rpm . . . . actual timing at 34.4* . . . . . coil fires and triggers the pulse which lasts the same 1 ms until it's killed by the 20* infrared reflector sensor.

Looking at 3600 rpm, with a radical timing of 41.6* . . . . the coil fires and starts the pulse, which then gets killed by the 20* infrared reflector sensor in . . . you guessed it, the same 1 ms as in the previous examples.

Same pulse width, (which seems to be what you're measuring) for widely varying timings just because of the rpm involved . . .

Isn't the meter going to read the same for each of those examples?

Help me out here . . What did I miss?

 

[PotvinGuy]

Duty cycle, grasshopper. Don't feel bad, I step away for a day and get confused myself. I gotta keep better notes. We use pulse width in degrees, not ms. Time and RPM are irrelevant, we are only interested in duty cycle, which is proportional to advance.

See that while the pulse is 1ms in both cases, the duty cycle changes. And the meter averages that. Now when I set it up on the bench I noticed at real low rpm the meter pointer would flicker. And the meter always made a little buzzing, poor thing. So I inserted an active low-pass filter ahead of it.

 

[Island Girl]

And the electromechanical nature of the meter movement effectively integrates the pulses?

How accurate does that turn out to be, and over what rpm range?

I can see it working under the right conditions, but curious about range and accuracy where you would want somewhere close to 2*.

 

[PotvinGuy]

This filter does the integration:

Works 600-6000 rpm with negligible error. The resolution and accuracy of the meter itself are the limiting factors.

 

[Island Girl]

This filter does the integration:

Works 600-6000 rpm with negligible error. The resolution and accuracy of the meter itself are the limiting factors.

Ahhhhh, the filter, (series RC with unity gain follower) don't recall seeing it in the linked paper, but it would do the integration . . .

How fast is the reaction time, particularly coming from lots of timing to much less timing?

 

[PotvinGuy]

IG, the meter seems to promptly reflect the timing as I drive; haven't noticed any settling time. Haven't done any quantitative testing. I don't even have a good schematic as built. Just threw it together.

 

[Guy]

I dont think you want to run without a thermostat the water will go though too fast and wont cool try a 160

 

[Indycars]

If it goes thru the radiator slower, then doesn't it have to go thru the engine slower,
picking up more heat ?

 

[PotvinGuy]

Read this:
Cooling secrets