gregeast

Cool stuff as always Rogue

bebbetufs

Great.
I have a Blaster coil sitting on the shelf, should I install it or do you expect to come up with something even better in the future than what you have already mentioned in here?

Rogue_Ant

Yes. As I make a bit more progress I will release more info.
It is going to be hard to beat the HVC-II coil though...

Dave W.

OK so what's the correct conversion? I thought it was X/2.5, but maybe the units are off?

Rogue_Ant

Dave, the conversion is not a static function - it has to take into account RPM, and is different depending on the model.

Rogue_Ant

Ok, so I tested the IGN-1 coil (the LS-style coil without built-in driver), and the D585 coil (which is supposed to be the most powerful version):



The interesting things about the LS-coils.
First I measured the IGN-1 coil to have 4.25mH of inductance. Which at 9 amps of current:

1/2 * 0.00425 * 9^2 = 172mJ

Which is pretty good! Since these coils are designed for one coil per cylinder, the coil has a significantly long time to charge; it follows that they designed the coil with a lot of inductance, which gives more potential, but slows down how fast it charges.
In-fact the coil is actually slower charging than the stock Bosch 944/951 coil - but since the coil has 4x the amount of possible charging time avilable, the slower charging is not an issue!

The D585 coil is even more interesting. First, it has a built-in driver with external heat-sink! Thanks to newer, better transistors, the built-in driver most likely has lower internal resistance than the factory DME driver. This means the coil windings see more voltage (since there is less drop across the driver transistor), which translates into a faster-charging coil.
The D585 with its built-in driver does charge the coil faster than the (similar) coil with the DME driver. That said - it does something very interesting if you push the dwell time (14volts):



You can see the coil fires before the dwell logic signal is finished...
This was a little surprising - I expected to see the coil do current limiting/holding once the amps reached a certain point, but I did not expect the coil to prematurely fire on its own. So, I changed the system voltage down to 11 volts, and increased the dwell time:



Again, we see the coil prematurely fire! And this also confirms that the coil will automatically discharge if current gets around ~8 - 8.5 amps; the coils internal threshold for automatically firing is not associated with the dwell, but only the current.
(For those keeping up with the math - this means the coil only has ~130-150mJ of energy, not the 170+ mentioned earlier.)

Well, this might not mean much to 95% of 944/951 owners, but we do have enough LS-swaps (and stand-alone EMS) which use these coils. The coils prematurely firing means they are firing earlier in time, which is the same as increasing ignition advance. So - be very careful about your ignition dwell settings with this coil. The coil dwell information on the D585 coils on the Megasquirt site, http://www.megamanual.com/seq/coils.htm, is not correct the D585 coils.
And just a little bit of proof, they state that the coil limits the dwell to no more than 8mS. Here is a screen of the D585 coil charging for well over 15mS of dwell (I changed the base-time divisions from the previous screen-shots, and ran lower system voltage):




I have quite a few more coils to test (coil-pack / wasted-spark). As I get those tested, I will publish the results!

Thom

If dwell has no influence on when a D585 coil will fire, why should dwell actually matter at all in setting up this coil?

Rogue_Ant

In order for the computer to accurately time the ignition event, it must have a direct way of firing the coil. In the D585 coil case, that is the falling edge of the dwell logic pulse. This is a correct ignition event for the same coil:



You see that as soon as the dwell logic goes from high to low, the coil fires. This is correct, and is exactly when the ECU desired. The issue with the D585 coil is that it will fire on its own if the current is too high.

What happens when the D585 coil builds too much current and it prematurely fires? Well, lets say you have programmed in 5.5mS of dwell into your ECU. And lets say you want 30° of ignition advance BTDC. The ECU must start charging the coil 5.5mS before 30°BTDC. So the ECU is predicting where the engine will be in time in order to charge the coil, and then fire it exactly the requested 30°BTDC. Now, if the coil fires prematurely, then it is firing earlier in time, therefore the ignition will be more advanced than desired.

I.E. If the engine is running at 6000RPM, the ECU is programmed for 5.5mS of dwell, but the coil fires early at 5.0mS (early by 0.5mS); this will increase ignition advance by 18° before the ECU expected 30°, giving the engine a total ignition timing advance of 48°!

It doesn't take a leap to see how this could cause extreme knock, or something more similar to pre-ignition - possibly hurting the engine.

The answer for those using this coil is to be 100% sure that your dwell settings are correct - even erring to the conservative.

Thom

Sorry for being thick (electrics in general and me have never been good friends) but I can't see on the two first graphs the difference that shows that you reduced current from 14V to 11V.

How did you figure out the "correct" 3.99ms of dwell for the last graph? Dichotomic hit or miss? What is the current going through the coil on this last graph?

Adker

There is an internal current limiting feature in the LS2 coils.

URG8RB8

If we don't don't use this type of coil with a built in driver, we will never experience this problem, but always like learning new things! Thanks for all the new and ineresting information generated by this thread Josh!

Josh:

Can you shed some light onto the table in the MS forum about "%age% versus voltage" in the Dwell Battery Adjustment screen and then the Chevy Silverado lookup table that follows? I am pretty sure this must be a built in feature of the Megasquirt software, but would like to hear your opinion.

I am wondering how this might affect the COP system on my FL car? Maybe not at all since they don't have built in drivers, but need to verify. I am not exactly clear how the MSD DIS-4 interacts with the Wolf software. Guess I will be doing some more research.

Rogue_Ant

I don't want to get too far away from the 944/951... Just understand that the D585 coils will fire at an unwanted (advanced time) if they build too much current...


The MS system uses a base dwell value, then one single-axis table which then scales the base value accordingly. This is a limited but working method of setting dwell.
The Sliverado table is how the DME does it. It defines a two-axis table, RPM vs Volts, and then grabs the appropriate dwell value. The major advantage of this is the ability to adjust dwell according to both voltage and RPM (very important if your coil charge is time-limited as RPMs increase).

Your FL car uses capacitive discharge, so most of this doesn't apply. (I went over capacitive discharge earlier in the thread)

URG8RB8

That is what I thought, just wanted to get your opinion. I ran track and went to a lot of great football games getting my BSME at UF! Plus that was a long time ago, and now living in Bangkok corrupts the mind!

Gentlemen, trust me on this, if you are single and have never been to Thailand do yourself a favor and book a ticket for your next vacation!

After two years now, I am looking forward to coming back and doing a little wrenching!

TurboTommy

Josh;
I've been toying with the idea of replacing the stock coil with the MSD blaster.
So, I checked their website for the right model number ,pricing, etc, and I happen to see that they have stated 8 millihenries of inductance in the specs. Calling their tech line comfirmed this.
You mentioned the inductance of this coil is 4.36 millihenries. Any idea as to the discrepancy. Does it even matter? (Except you mentioned that generally the higher the inductance the higher the likelyhood of longer charge times)

Rogue_Ant

The discrepancy might be due to how they are stating the inductance...
Since an automotive coil is actually two inductors, they might be taking the average or function of the primary and secondary winding values. The primary inductance is what has an affect on coil charge rate, and the primary was the value I measured.

Regardless, I wouldn't worry about the difference between my measurement and the stated one.